Contents

• Chapter 1Why This Book
• Chapter 2The Demographics Are Changing
• Chapter 3When Tech Shifts, The Work Shifts
• Chapter 4What Ranking Means To Your Client
• Chapter 5The Need For An Index
• Chapter 6How You Can Use The “Standard” Ranking Factors
• Chapter 7Where To Start Working
• Chapter 8How to Work the Factors
• Chapter 9Free Resources To Help Rank Your Client
• About the AuthorAbout the Author
• Appendix AWhat the Categories Mean
• Appendix BThe Full Factor List

Why I wrote this book

I'm going to start out this book in an unconventional way because I want you to understand why I even wrote it. I didn't need to write it because I needed to be more famous, or wealthier and it sure wasn't because I was bored. I wrote it because someone had to. The 'ranking industry' needed a kick in the pants to stop arguing about what ranked businesses and to start really working on it.

Let me give you an idea of why I didn't even want another big project. I semi-retired about 10 years ago and I've been living in the coffee triangle of Colombia and I can literally walk five minutes and pick coffee cherries off the plants. I spent a lot of time in New York, Seattle, San Francisco and DC in their best coffee shops with friends and associates and I remember the conversation going like this. "I wonder what coffee tastes like right off the farm, freshly dried and then roasted". Well, now I know. And I was set down here because besides this home in the highlands I have another one on the coast. I can throw a rock from my balcony and watch it sink in the Caribbean Sea. It's in a remote area and most of the fish have never seen a fish hook and the lobsters there are as big as your forearm. So, I had no desire to head back to the US and join the rat race again.

But AI happened—and I saw it was going to destroy a lot of the businesses that millions of good people worked for years and years to build.

As I got more and more involved in figuring out what needed to be done to help these businesses, I realized there was another group that was going to get hit equally hard—an industry I had been passionate about for over 40 years. Online marketing, especially SEO, was going to be under a lot of pressure. Neither group did anything to cause this—it was just another seismic shift in technology that had caused similar disruptions before. I knew I could never reach the millions of individual businesses but I could help the marketing industry that was sure to pivot to helping them get ranked in AI.

The Paradigm Shift I’m Hoping For

Let me start off with a story that I would never have told anyone years ago in my 30s and 40s, when I was in business. In fact I did everything I could to hide the little bit of a Texas accent that I have when I speak. I have been in boardrooms from coast to coast, and I don't think anybody really recognized what my background was.

When I was a teenager, I wanted to be popular like any other kid, and one of the most popular kids in our school was of all things—a bull rider. One day I asked him if he would teach me how to ride as well, and he just stood there and chuckled lightly, then said, "Ok, meet me at the rodeo this weekend behind the gates." I don't think he really expected me to show up, but I did, and when I dropped down on top of that 2,000 pound sweaty piece of quivering meat, I should have been afraid, but instead I was eager to give it a try.

My friend saw that I was serious about this, and we started going to rodeos almost every weekend together. I didn't realize it at the time, but I was extremely lucky to have him as my mentor. He was very serious about it, he was very good at it, and he went on to become one of the top rodeo professionals in the nation.

I was ok as a rider, but not great. It was not until my parents made me go to a bull riding school that things changed for me. The school was a full week long and they gave us a lot of new information, but of all the things I learned one thing changed who I was.

Three words: "Watch his neck".

It allowed me to go from being like teflon, sliding off most of the bulls I tried to ride, to super-glue-and-velcro. Out of hundreds of bulls over the next two years, I might have been thrown off a dozen. I had the most consistent bull riding record in Texas my senior year of high school and won the TYRA bull riding championship, gaining my friend and I the title of best two bull riders in Texas. The problem I had and why most people don't ride is they watch the bull's head to see where he's going next. They fall off when they shift their weight based on where the head's going.

This was the first time that I had been exposed to a "catalyst" like this. It was almost like that tiny new "insight" was magic. It changed me: I got a confidence and swagger I'd never experienced before.

The same thing happened to me a few years later. My dad owned an insurance and real estate company in our small town of El Campo, Texas, and he got lung cancer and asked me to drop out of college and come home to run the businesses while he either passed or got better. Luckily he got better, but we never really agreed on how to market the company. At just 19 or 20, I was afraid to go out and shake hands with 60 or 70 year old farmers and business owners that needed to buy insurance from me…so I came up with other ways to bring in customers using sales letters, postcards, doorhangers, etc. It was working, but it wasn't his way.

So, I moved an hour away to Houston and set up my own insurance agency.

I quickly discovered getting started was expensive. By the time I paid for office space and employees and everything else, the bank account was almost empty. I knew a few months down the road I'd be completely broke.

One day, pondering my misfortune, I happened to pick up a copy of Inc. magazine and it had an article about a man named W. Edwards Deming. Mr. Deming is now known as the father of quality in Japan and was largely responsible for turning around their manufacturing sector after World War II. His quality engineering process is much more complex than this, but what I got out of the article was try, measure, repeat, the “Quality Loop.”

Deming's process applied to car design, but I didn't know any better and thought it would also work for my agency. So I started improving everything (and measuring). The new "Quality Loop" was applied to my newspaper ads, my postcards, my brochures, my handouts, what we said on the phone, how we closed calls and every customer interaction was measured for response and tracked on the whiteboard in my office. And, it actually worked and we started growing fast, insanely fast. In fact, about 18 months after opening, I was featured by the Houston Independent Insurance Association as the 'Fastest Growing Insurance Company in Houston, Texas' out of over 7,000 other insurance agents.

I was about 23 at the time and was loving life. I had just found another 'catalyst' that helped me find my 'swagger' again.

You guys are marketers, and that formula doesn't seem much like magic today — it's more standard procedure, right? But to me back then, it was another catalyst that changed my fortunes again.

One more quick story and we'll get into the meat of the book. There was a man named Rich Harshaw that owned Monopolize Your Marketplace, one of the fastest growing marketing agencies back in the 80's. As a lead generation tool, he had created two audio files telling business owners how to adjust their marketing. He cited over a dozen client stories that demonstrated how powerful his method was, but the one that got my attention was a client family that had immigrated to the United States with nothing, that spent a year working for other moving companies, and then bought their own trucks. Then they did what most of the other moving companies did which was to pay $3,000 or more a month for their yellow pages ad and sit by the phone. They made their ad like the other moving companies, and got a whopping 70 calls a month which kept them just above bankruptcy and starving. The second year, after getting a copy of the audio program and hiring one of Rich's guys to rewrite their ad, they got different results. The first month the new yellow pages book came out, they got 955 calls.

Yes, I had to double check that number too.

I was hooked and amazed. His method taught business owners back then to present their content as if you were drinking a beer or on a camping trip with a good friend. I heard 'find the heart' in his method, and that has allowed me to essentially turn an average campaign into one that could 10X sales. That was like being given the secret to the universe.

And again, I had proof that a tiny phrase, a small change in a process or how you placed words on a page could be the 'catalyst' that can separate you from struggle and give you a new level of confidence and a swagger that literally "goes before you" in life.

Nearly every creative I touched after that shot up conversion rates — I was having so much fun that I sold my agency and have been either in marketing or technology ever since. That catalyst enabled me to generate hundreds of millions in revenues, and explosive growth for most of the companies that I've started, worked for, consulted with or owned a part of.

Building the systems and analyzing the data behind this book was not a first time project for me. I had spent almost 15 years working with data warehouses, data marts, etc., in lead positions responsible for reverse-engineering and other systems projects in some of the largest government and private companies. Setting up the environment so that it was 'free of opinion' from my team, myself and our new AI friends was the first thing I did.

I’m hoping that what I present in this book will be a paradigm shift for the professionals that are trying to move into the new “ranking industry”.

I think everyone that reads the book will be surprised at our findings and how valuable a new ‘Google-like’ index and checklist would be.

What I hope this book does for you

So, while this book might be dry, analytical, controversial and even adversarial at times—I'm hoping you'll find it as a catalyst. Something that will encourage you to get out of the argument about what AI needs and leapfrog over the experts, consultants and agencies that are stuck.

In fact, I’m hoping the three words “Adopt the Standard” will stick with you long after you set this book down.

Who the book is for, and not for

I published our internal foundational factor list to help as many businesses and marketers as possible 'get started'.

Two kinds of people will get the most out of it. The first is the agency owner, or the person who does the SEO. Their clients have started asking a question the old search methods can't answer: why does the AI never name us? The second is the marketer or web lead at a local business. They can already see that being recommended by a machine is about to matter more than ranking on a page, and they want the actual checklist instead of one more essay about it.

This book is not for national or international brands, large franchises, or multi-location businesses. They already have thousands of reviews and do not need the foundational factors to get recommended. In a franchise, the main brand domain carries about 60% of the weight. A single location can do little to change its own score, even with its own subsite. This book is for the agencies, consultants or operators with clients that operate from a physical location or one local service area. We may write a guide for the big brands later.

Search changed shape, from a long list you choose from, to a short list of 2 or 3.

Technology has shifted consumer preferences again

For twenty years the job was to rank on a page. There were ten blue links and a map pack, and you scrolled and you clicked. A business competed for a spot, and landing on page one meant you were at least in the running. That world is quietly ending, without a headline to mark it. A person asks an AI assistant — ChatGPT, Gemini, Google's AI Overviews, Perplexity — "who's the best plumber near me." They don't get ten options to sort through. They get two or three names, already picked for them.

The market moved from search to recommendation. The choice used to belong to the customer. Now it happens inside the AI assistant, before the customer sees anything. A business on page one used to get a fair chance at being picked by a person who was looking. That chance is gone. If the AI assistant does not name the business, the customer never learns it exists.

Sources: Gartner, “Search Engine Volume Will Drop 25% by 2026 Due to AI Chatbots” (Feb 2024); McKinsey & Company, “New Front Door to the Internet: Winning in the Age of AI Search” (Oct 2025); Search Engine Land, “37% of Consumers Start Searches with AI Instead of Google” (2025); Dataslayer, “Google AI Overviews: The End of Traditional CTR” (2025); Exposure Ninja, “AI Search Statistics for 2026” (2026); BrightLocal, “Local Consumer Review Survey 2026” (2026); Click Intelligence, “Sundar Pichai Says Google Search Will Change Profoundly in 2025” (2025).

The Death Cross

Traditional search is falling while AI discovery is climbing, and the two lines meet around 2026, the point where about half of the searching has crossed over to AI and does not cross back. That is the death cross. The numbers behind it are not subtle. In 2024, Gartner forecast that traditional search volume would fall 25% by 2026. The measured numbers since have gone further. By October 2025, McKinsey found half of all consumers already using AI search, and put the traffic decline facing brands that don't adapt at 20 to 50 percent. Thirty-seven percent of consumers now start a search with AI instead of Google. When Google runs an AI Overview, organic click-through falls 61 percent and paid falls 68 percent, and six in ten searches now end with no click to any website, eight in ten when an Overview is on the page. For local businesses the shift is faster still: in a single year, the share of consumers using AI to find one rose from 6 percent to 45 percent, while Google's share of that job fell from 83 percent to 71 percent.

Examples from the past

This is not the first time technology changed what people use. It happens again and again. Each time, the businesses tied to the old way lose.

The PC replaced the typewriter. The fax replaced the mailed letter, and then email replaced the fax. Recorded music went from the LP to the 8-track to the cassette to the CD, and then to the download and the stream. Movies went from the theater to the VHS tape to the DVD, and then to streaming.

Blockbuster had more than 9,000 stores and did not move to streaming. Netflix did. Blockbuster is gone.

Chegg didn't see it coming. Students paid Chegg for homework help and textbook answers. ChatGPT launched in November 2022 and gave the same answers for free. On May 2, 2023, Chegg’s stock fell 48% in a single day, after its CEO said ChatGPT was cutting into new customers. Its subscribers dropped from 7.8 million to 3.2 million, and its market value fell from $14.7 billion to about $115 million, a 99% loss in three years.

There are roughly 36 million businesses in the United States. Several sources confirm that more than 18 million of them operate from a physical location and/or serve a local area. Since we’re seeing AI recommend only 5% to 15% of businesses when it refers a new prospect, that splits the local universe into the 1 to 2 million it names and the 16 million it doesn’t. If AI does a full sweep the way other technologies have — DVD to streaming, newspapers to video feeds — as many as 16 million businesses could literally go the way of Blockbuster’s almost 10,000 stores. This is a HUGE OPPORTUNITY for anyone who chooses to become an expert in this field, because practically every business is asking the same question: “How do I get ranked in AI?”

Whether the rest of the world sees it yet or not, we're at the front-end of the biggest technological shift in the last half century. This is when fortunes are made and lost.

The current debate

Normally I would write a book or report on a topic and stay out of the debate, out of anything controversial, because I didn't want it getting in the way of the point I was trying to make. But this time it IS the point. So you're going to find a steady comparison of SEO and other forms of marketing to what AI is forcing us all to do. Regardless of which side you're on right now, I hope this book gives you something to help you transition — from whatever title you use to describe your expertise, to whatever title this disruption is going to cause you to be on the other side.

Every agency that used to sell "SEO" now has a new service to sell, but from what we gather from the forums, Reddit, and other groups where the industry is discussing what to do, we're seeing a lot of confusion, misinformation, and conjecture expressed as facts. Nearly every large marketer is scrambling to fit in — but to date none have proven they have the magic formula to consistently rank their clients. Pity the poor client who trusts someone to advise them whose only tools are a visibility tool using AI, AI itself, and the standard SEO tool suite.

The current chaos is no different from back when Google was number three, behind Yahoo and AltaVista. I was working for a company then, a dot-com startup that had just raised $100 million from investors like Dave Thomas of Wendy's, Jim Patterson of Long John Silver's, and Governor Wallace of Kentucky. As their Director of Internet Marketing, there was a lot of pressure on me to get our company and our products listed at the top of these three search engines. There was no book, no guidelines, no classes, and no help from the companies themselves. It was just a small team and I, and the 15 years of business process reengineering that I had before taking this position.

We're not the only ones who noticed. The whole search industry is in motion right now, and it's worth pulling that conversation out into the open where you can see it.

And there are no classes now, no training, no help from the inside of the AI companies. Same as before, the right way to identify how to rank your client's business is to deconstruct the top ranking businesses. If anyone else has done it, they're not sharing the information.

Two years ago the debate was whether AI search even mattered. That argument is over. Gartner's 25%-by-2026 call was controversial once. Now it reads like a complete miss next to the real data. The big agencies and marketing platforms have stopped asking whether to adapt and started publishing their own AEO and GEO playbooks.

Everyone is scrambling to figure out what 'actually works' to help their clients.

Is SEO the same as AI Ranking?

I've tried to make this book as neutral as possible. There is a lot of debate across the entire marketing industry about what AI actually requires to rank. On one hand, there are the SEO purists who say ranking in AI is exactly like SEO. Others believe it's an entirely new discipline. I'm not taking sides, but having witnessed both startups, Google and now AI, I think you should keep this image in mind. Imagine a class being taught by a teacher who for 20 years has rewarded their students for keeping their head down and following a long list of rules. Then without asking for permission, six unruly and hyper-intelligent kids take over the class and could care less about the previous rules, and in fact will lie if you ask them what the new rules are.

What are the similarities

Much of the on-site work is the same. Titles, headings, schema, page speed, a secure HTTPS connection, a clean mobile layout: those are SEO factors, and they are on the factor list here too. Content analysis is the same. Proving real expertise, answering the questions a customer asks, and writing clear service pages all carry over. And where SEO targeted keywords, AI ranking targets prompts. 'Plumber near me,' 'emergency Sunday plumber,' 'best running shoes for flat feet': the local, service, and product prompts a person types into an AI assistant are the same intent an SEO used to chase with keywords.

For the nimble marketing agency or SEO expert, they are not losing a job, they are just losing a job title.

What are the differences

The fact is, while many factors are the same, the weighting on those factors shown to matter to AIs varies widely. Search engine optimization weighed heavily on keywords and backlinks. What we are seeing weighted now is identification, reputation, and corroboration.

No SEO guy ever asked a client to link their medical certification back to the school's website so the consumer could check it. No SEO expert ever suggested that a restaurant link to the local newspaper's coverage of the graduation ceremony of its three new chefs, and then ask the author to make sure the restaurant's website was linkable from the article. But the skills are similar, even though the strategies are changing drastically.

There is no mutually agreed upon index of web pages; in fact, AI doesn't need to go to Google at all. It's already absorbed Google and close to all the pages on the planet, and practically the contents of every book or publication as well, plus it's pulled in Dun & Bradstreet, the BBB, Hoovers, and every business directory that matters. It could basically tell you about a business without ever hitting Google.

One of the reasons the debate continues is that the information the industry is getting from the AI apps themselves is total bunk. The AI apps are not going to tell you how to rank them, at a granular level, because the actual chat or API that you're asking the question to isn't privy to the exact program that's being used.

You can ask one of Tesla's new robots, 'What program do you use to lift your arm?', but you won't get the 'right' answer, because only the programmers have that information. Too many people expect the AI to give them its source code. Only the individual AI companies have it, and so far none of them are as forthcoming as Google was with its community.

I expect the argument will rage on for a few more years.

The 3 list vs the 3 billion list

Google returned a list. It might hold billions of pages, but every ranked page held a position in the index somewhere. A business could see exactly where it stood, page one or page ten, position four, and measure progress week to week.

An AI assistant returns three names. Sometimes two. There is no list to find yourself on. A business is either one of the three or it is nowhere, and nowhere gives no signal. There is no way to tell if a business is the fourth choice or the forty-thousandth.

Giving the industry a way to rank and score a business on an 'index of sorts' that is beyond the nebulous mentions that most people are using now, is what I aim to do in this book.

Here's a fair question every client asks: if you get me into one of the top spots in AI, what is that worth?

I'll be honest with you. I have not found a way to calculate it exactly, and most of us sitting across from clients are in the same spot. What I can give you is the rough way we lay out the possible upside, so an owner can at least see the shape of it. Every number here is a projection, and the right move is to adjust it to whatever real data you can find.

It starts with where the money comes from: new prospects each month, the brand-new customers, not the repeat ones.

How We Show It

We start with the crossover graph from the front of this book. We show the client that at the end of 2025, a business like theirs was pulling in about $5,000 a month from brand-new customers. For many businesses that is a low number, because customers are always leaving — moving away, passing on, changing what they want — and new ones have to replace them just to stay even.

Then if we look around the neighborhood, the competitors on Google, local directories, etc., we could easily find 50 competitors for most businesses in any metropolitan area. Obviously every geographic area and every business type could be above or below that threshold, but for ease of calculation we use 50. At $5,000 each, they are splitting about $250,000 a month. For most businesses at the end of 2025, most of that came from natural search, local search, and word of mouth, usually a referral from someone the customer already knew.

Since 2025 it has shifted fast. Today, roughly half of all new prospects come through AI, whether that is an AI answer directly or a referral that began with someone asking an app. That moves half of the $250,000, about $125,000 a month, toward the businesses AI actually recommends.

And AI does not recommend many. The stats put it under 10% of local businesses; to stay conservative, let's say 20% win its attention. Twenty percent of 50 is 10. So ten businesses now split that $125,000, which is $12,500 each in new AI money. The non-AI half that is left is spread thin across the whole field, so every business's baseline is cut in half, to about $2,500 a month. The forty AI never names are stuck with just that base, down from $5,000. The ten it does name keep the $2,500 and add the $12,500, coming out to about $15,000 a month each.

It does not stop there. By the end of 2027, and the sources we have found agree on the direction, we expect about 75% of new-prospect referrals to come through AI. The same math makes the gap brutal. The businesses AI ignores drop to around $1,250 a month, a quarter of the $5,000 they started with, while the ten AI keeps recommending are splitting most of the pie, on the order of $20,000 a month each.

The 10 AI recommendsEveryone else

Is this exact? No. It is a projection. Use whatever real numbers you can gather for a given town and trade. The result is the same either way. A few businesses take almost everything, and the rest get very little.

The hard part isn't the math. It is that most owners still don't believe this is happening, or that it could happen to them. I believe those that deny this is happening and don't start to act will be the hardest hit.

If you're reading this and you've found a better way to show a client what's at stake, I'd genuinely like to hear it. Reach out.

What the index does

Having an index like the one search engines have gave marketers a milestone, a starting point, a goal. Mentions are not at all the same. They are more fluid, and while they can be counted, they are everchanging, almost on a daily basis. Then there are the hallucinations, and the lack of timely training-data updates. And there are five or six hyper-competitive companies all jockeying for top position. AI has no publishing cycle like Google's algorithms, and it doesn't appear to care at all about giving the new 'ranking experts' a real answer as to why one business ranked over another.

In the first month, I used every AI and every visibility tool I could find, until the inconsistencies in where clients ranked, the outright hallucinations, and the red herrings I got from AI made it impossible to continue successfully.

That's when we decided to build a system to extract the index. Even if I couldn't have a search-engine-style view of every ranked prompt across the world, I could at least build a 'shadow index' of the factors that were prominent in the majority of AI-recommended businesses.

The problem I kept having is that AI can code, and it can reorganize data, but it could not give me a top-to-bottom list of all the factors that I needed to rank client businesses. It kept leaving out the foundational issues, and depending on the engine, the day, or the month, it would only keep half a dozen 'tips to ranking in AI' consistent enough to consider them 'real'.

Someone gave me this example the other day. He said, in many ways AI are like humans, and if you ask a father what things are important that he does for his daughter, he'll give you a list of 10 things, like tell her I love her, teach her about boys, go to her sporting events, etc. But that's really a partial list because the foundational things get overlooked whether you ask a human or AI. What he should have said is, I protect her, above all else. I make sure she has water, food, medical care, good grooming, and education.

I knew that there was not going to be a huge single index, probably ever again, but I also knew from my experience that I could build a shadow list of all the factors that the top-ranked AI-recommended businesses exhibited. If I could find out the factors that are present, then I could build the checklist that had served the Google ranking experts for so long.

Just like if you ask AI for a good bakery near you, you get an answer, and even the criteria. It does not tell you it only recommends a business with a front door and glass in the windows. That sounds ridiculous, but it is true. A business without the foundational things, the front door, the glass, the electricity, does not get recommended. There is no point debating whether citations, mentions, listicles, or reviews matter more, until the foundational factors are in place.

Regardless of how you ask, the AIs will skip ranking factors.

What Is Required To Rank This Bakery
Ranking Factors	What AI Says	What AI Forgets
Good Reviews	✓
Clear Description of Services	✓
Google Profile	✓
Off-Site References	✓
Roof on Building		✓
Electricity		✓
Front Door		✓
Glass in Windows		✓
Proper Schema		✓
Specific Title Layout		✓

Fictitious example only. Every check in the red column is a factor AI never mentioned but should have.

Were there factors that AI was not giving us that mattered to the ranking process? We knew there were, and there was a way to deduce what they were.

No matter how good the strategies or tactics that a person is using to rank a business, its services, or its products in AI, there are foundational factors that need to be resolved before the full 'lift' is possible.

What it would look like

Before we built the systems to find those foundational factors, we had to picture what the output would look like and how it would be used. We imagined there would be about 250 things we could possibly test any business for. We identified seven categories of inputs: the rendered website, the network layer beneath it, the domain registration record, the performance measurements, the outside profiles like Google and the review sites, the independent sources that could corroborate a claim, and a structured read of the page by an AI. We were not sure we could get to all of that data, or that once we had it, we could test it reliably.

And so that is what you're going to be looking at today, a result set of roughly 250 tests that over time distilled down to the 141 criteria that are most often present in top AI-recommended businesses. Not every business will have all of them present, in place, or aligned. In fact, we found that the average top-ranked business consistently has about 100 of them. The average unranked business has about 75 or 80 of them aligned.

From there, the score is simple arithmetic. Whether we start a scan ourselves or a business runs a free one at AIBusinessScore.com, the data flows into the same system, and the score is just the share of those 141 baseline factors the business has in place. The work never really stops: the analysis and the prominence ranking re-run on their own every day, and any new domain that enters the system gets scanned the same way.

Keep The Data Free From Opinion

The last thing I want to say before we get started is that the data that I'm going to present to you is not my opinion. It's not filled with my personal interpretation or my team's influence.

The standard that I'm going to give you today was not us asking AI how to rank businesses, or how to rank products or services. The standard is the work of our system literally evaluating hundreds and hundreds of factors, most of which can be extracted from either websites or external data, until the data revealed a set of criteria that exists on the majority of top-ranked local businesses.

Let me define what we did to get to our conclusions because "reverse-engineering" can have different meanings depending on where it's applied. We started with the universe of things you can test a business, domain, and website for.

We asked the data, math and science to get the answers being careful to not let AI or human opinion influence the outcome. This empirical data set is what I believe should be the key factors that professionals in the new AEO/GEO industry should focus on first, or at least in parallel with the other known factors such as content relevance, to establish a foundation.

Building the evaluators

We started out to answer this one question: what do the businesses AI already recommends have in common? To answer it, we drilled into their websites with every method we could, and pulled their off-site signals too. Once the collectors were feeding us clean data, we built an evaluator for each thing worth testing.

At the peak we were running almost 250 separate tests on every AI-ranked business. After roughly 10,000 analyses, the list settled at around 150 factors that showed up in most AI-recommended businesses, and we have since tightened it to 141. It has held steady at 141 since March of this year.

The engine we built to distill down these factors started as the backend to AIBusinessScore.com, our consumer-side website that allows businesses to see what they need to do to 'fix' their businesses to prepare for ranking in AI. When we first started, I had an uncomfortable feeling about providing advice that I could not definitively prove. I'm sure some of you are advising clients or stakeholders and feel the same uneasiness.

In order to get the best answers possible, we ramped up our technical resources. There were many false starts, and many factors we felt just had to be influential to AI's decision that the data proved were not. Always the data, always the facts. It took us about 10 months, but now we've proven with mathematical certainty that the foundational factors in this list exist in most highly ranked AI-recommended businesses. It did not take us 10 months to scan the 30,000 businesses. It took us ten months to develop the collectors, evaluators, results interpreters, and the reporting necessary to be sure the result sets were valid.

Where the data comes from

Every score comes from the public information about a business, the same information a machine would use: its website, its Google profile, its reviews, its listings, its domain record, and the public data tied to its name. We don't ask the business anything. For each business we run seven passes, and each one pulls a different kind of information.

The rendered site. The first and largest pass is the business's own website. We don't just grab the raw HTML; we load and render each page the way a real browser would, then read what is actually on the screen. From that we pull the name, phone, address, and hours; whether there are real about and contact pages; the title and meta tags; the headings and structure; the schema markup; the images and their alt text; the links, buttons, and social profiles; the policies; the trust language and licensing claims; and the quality and specificity of the content. This one pass feeds more than half of all the factors. It also tells us something most tools miss: whether the content even survives with JavaScript turned off, which is close to what a machine sees.

The headers and the network. The second pass works below the page, at the network level. It reads the site's response headers and its root files. From here we check the robots.txt file — including a separate check for each individual AI crawler it might allow or block — plus the sitemap and the llms.txt file; the security headers and the SSL certificate; the response speed, redirects, and page size; the DNS records; and how the site handles a page that isn't there.

The domain record. A quick lookup of the domain's registration tells us how old it is and how long it is committed to stay. Two factors, but they carry weight, because an established domain reads differently than one registered last month.

The performance measurement. A separate pass measures how the site actually performs, on both mobile and desktop: its speed scores, its core loading metrics, whether its images are optimized and its code compressed, and whether anything is blocking the page from painting quickly.

The outside profile. Another pass pulls what the wider world already publishes about the business: its main profile listing, its reviews, and its directory citations. From that we look at whether the profile exists and is verified, how complete it is, its category and photos and hours; the volume, freshness, spread, and responsiveness of the reviews; the directory citations; and, importantly, whether the name, phone, and address on that profile match the ones on the site.

The outside corroboration. This pass checks the business against independent sources, to see whether its story holds up away from its own pages. It looks at whether the business is recognized at all, what it is known for, its reputation, the year it was established, its presence on outside review and rating sources, and whether claims like a professional license actually check out.

The AI read. The last pass is a QA read of what the evaluators returned. This AI step cannot change the data or massage it or touch it. Its only role in the process is to double check that data exists where data should exist. It's worth repeating that AI is not involved at all in the prominence scoring, and our numbers and conclusions are simple math, a series of pass-fail criteria.

A few factors deliberately draw on more than one of these passes. The clearest are the consistency checks — they exist precisely to compare what a business says on its own site against what its outside profile says, and to flag it when the two do not match.

The Foundational Shadow Appears

Our tests were like an ultrasound that sends a spray of rays at a mother's stomach. The ones that pass through open space tell you nothing — it's the rays that bounce back off the baby that let you see the shape that's really there. We were doing the same thing, except picture us firing 250 lasers from the same spot at every top-ranked business, one after another. Pretty soon the shape appears.

Each of the Factors Fell Into One of 12 Categories

Here's the typical local business pass rate across the twelve categories of factors, which are weighted and shown in detail further along in the book.

Category	Avg. pass rate		Read
UX & Conversion		81%	Most pass
AI Visibility		77%	Mostly fine, but the failures are fatal
AI Corroboration		48%	Few have it everywhere it counts
SEO (technical)		43%	Basics pass, quality fails
Content & Expertise		41%	Wide open
Trust & Safety		37%	SSL yes, depth no
Business Identity		35%	The basics aren't on the page
Contact Prominence		35%	Reachability gaps
Local SEO & GBP		34%	Half never claimed or matched the profile
Social Presence		34%	Lightest-weighted
Service Trust (LSA)		30%	Service trades under-signal
Reputation		28%	Lowest scored — heavily weighted

As you can see, most of what every SEO and new AEO/GEO expert says lines up: these categories of factors are important to ranking in AI. This just further validated that our deductions were sound. In a future chapter we'll discuss the categories that carry the most weight, and what to work on as you align the foundational factors.

The Early Assumptions

What was surprising was that much of what the SEO experts said mattered did matter. But there was an underlying foundational issue no one was addressing.

It was the same issue the AIs were not addressing. There are key things that need to be in place before an AI will choose a business. If they are not aligned properly, the AI may be confused about a business's identity. If the site uses excessive JavaScript, the AI, which focuses on extracting the HTML, will miss key information. If the phone number is not clickable on mobile, that could have an impact. If the office hours on the GBP say one thing and the site says another, it could cause the AI to bypass a business. If the reviews show a negative trend or the responses are stale, that could cause AI to bounce off a site. And when there is no About Us page, no email, and no contact form, what happens?

We've gone factor by factor to understand the impact each one could have, and all 141 are important to a degree. There is a direct correlation between the number of factors and the position in the rankings. In almost every case, the more factors aligned or present for a business, the higher the ranking it achieves.

There is one exception. National brands, franchises, and multi-location businesses rank on the brand alone. They already carry thousands of reviews, so the factor count on a single location does not move them the way it moves a local business. In a franchise, the main brand domain carries about 60% of the weight, and one location can do little to change its own score. Everything here is for the local business.

The “Foundational” Alignment Factors

The ranking factors below are not sorted in any particular order, but each one is grouped into one of 12 categories. To display them here we’ve shortened each to its name; the full detail on every factor is in Appendix B.

Having this factor list allows us to score any domain and its competitors. This provides a reference point, or ‘standard,’ to measure against, and it also provides a deep list of factors to align to help your client rank. Again, this list was not derived by a human or AI’s opinion, and as of June 2026 shows the factors present on 30,000 AI-recommended businesses.

Along with the factor name and category, the table has a column for a client domain and three competitor domains.

The recommended use for the factor list is to check a client’s site first for the factors that are in place, then check the top competitors to that client. Once each factor either passes or fails, the percentage of passes out of the 141 determines the percentage score.

As you can see in this example, the client scores 58% and the closest competitor is at 64%. From here the agency knows exactly which factors to work on to catch up. The first goal is to start being recommended alongside the nearest competitor.

I believe this should be this industry's new checklist.

Ranking Factor	Client	Top Ranked A	Top Ranked B	Top Ranked C
Business Identity
Business Name In Title	✗	✗	✗	✗
Business Category Clarity	✗	✗	✗	✗
Years In Business Signal	✗	✗	✗	✓
Location Validation Strength	✗	✗	✓	✓
Team & Role Transparency	✗	✓	✓	✓
Domain Age Verification	✗	✓	✓	✓
Business Email Configured	✓	✓	✓	✓
About Page Present	✓	✓	✓	✓
Legal Entity Type Disclosed	✓	✓	✓	✓
Local SEO & GBP
GBP Listing Present	✗	✗	✗	✗
GBP Category Aligns With Site Content	✗	✗	✗	✗
GBP Core Profile Fields Complete	✗	✗	✗	✓
Photo Coverage Strength	✗	✗	✗	✓
Weekly Operating Hours Coverage	✗	✓	✓	✓
GBP Core Listing Fields Present	✗	✓	✓	✓
NAP Elements Present On Site	✓	✓	✓	✓
Contact Channel Readiness	✓	✓	✓	✓
Service Area Specificity	✓	✓	✓	✓
Map Embed Present	✓	✓	✓	✓
Keeps Prospects Informed (Google Posts)	✓	✓	✓	✓
Reputation
Review Volume Strength	✗	✗	✗	✗
Review Freshness Indicators	✗	✗	✗	✗
Review Sentiment Balance	✗	✗	✗	✗
Responsiveness to Reviews	✗	✗	✗	✗
Negative Review Pattern Recognition	✗	✗	✓	✓
GBP Photos And Reviews Present	✗	✓	✓	✓
Rating Distribution Health	✗	✓	✓	✓
Complaint History Signal	✗	✓	✓	✓
Known-For Clarity	✓	✓	✓	✓
Review Cadence	✓	✓	✓	✓
Award or Recognition Present	✓	✓	✓	✓
Service Trust (LSA)
Services Clearly Described	✗	✗	✗	✗
Contact Options Available	✗	✗	✓	✓
Licensing Signal On Site	✗	✓	✓	✓
Insurance Signal On Site	✗	✓	✓	✓
Business Owner Identity Signals	✓	✓	✓	✓
Local Profile Completeness	✓	✓	✓	✓
Contact Prominence
Email Address Present On Site	✗	✗	✗	✗
Contact Form Accessibility	✗	✗	✓	✓
Call-to-Action Buttons Present	✗	✓	✓	✓
Clickable Phone Tags	✓	✓	✓	✓
Contact Page Present	✓	✓	✓	✓
Physical Address Visibility	✓	✓	✓	✓
Operating Hours On Site	✓	✓	✓	✓
UX & Conversion
Above-the-Fold Clarity	✗	✗	✓	✓
Booking or Availability Signal	✓	✓	✓	✓
No Intrusive Popups	✓	✓	✓	✓
Call-to-Action Variety	✓	✓	✓	✓
SEO (technical)
Page Load Speed	✗	✗	✗	✗
Title Tag Quality	✗	✗	✗	✗
Image Optimization Indicators	✗	✗	✗	✗
Meta Description Quality	✗	✗	✗	✗
H1 Tag Presence	✗	✗	✗	✓
Script Efficiency Signals	✗	✗	✗	✓
Front-End Best Practices	✗	✗	✓	✓
Image Alt Text Coverage	✗	✓	✓	✓
Server Response Time	✗	✓	✓	✓
Internal Link Structure	✓	✓	✓	✓
Main Content Load Speed	✓	✓	✓	✓
Canonical Tag Presence	✓	✓	✓	✓
Responsive Layout Stability	✓	✓	✓	✓
Page Payload Weight	✓	✓	✓	✓
Sitemap Availability	✓	✓	✓	✓
Desktop Load Performance	✓	✓	✓	✓
Robots.txt Published	✓	✓	✓	✓
Breadcrumb Schema Present	✓	✓	✓	✓
Healthy Page Size	✓	✓	✓	✓
Schema Depth and Richness	✓	✓	✓	✓
Review Markup with Data	✓	✓	✓	✓
Low Redirect Count	✓	✓	✓	✓
Total Page Response Time	✓	✓	✓	✓
Site Reachable Without Errors	✓	✓	✓	✓
Viewport Meta Tag Present	✓	✓	✓	✓
Technical SEO Score	✓	✓	✓	✓
AI Visibility
Allows Anthropic ClaudeBot	✗	✗	✗	✗
Allows ChatGPT On-Demand (ChatGPT-User)	✗	✗	✗	✗
Allows Google-Extended	✗	✗	✗	✗
Allows OpenAI GPTBot	✗	✗	✓	✓
Allows OpenAI Search (OAI-SearchBot)	✗	✓	✓	✓
Allows Perplexity On-Demand (Perplexity-User)	✓	✓	✓	✓
Allows PerplexityBot	✓	✓	✓	✓
llms.txt File Present	✓	✓	✓	✓
Established Entity Recognition	✓	✓	✓	✓
Indexing Allowed	✓	✓	✓	✓
Crawler Access Allowed	✓	✓	✓	✓
LocalBusiness Schema Present	✓	✓	✓	✓
Reputation Standing	✓	✓	✓	✓
Open Graph Tags Present	✓	✓	✓	✓
Sitemap Linked In Robots.txt	✓	✓	✓	✓
Allows Claude On-Demand (Claude-User)	✓	✓	✓	✓
Allows Claude Search (Claude-SearchBot)	✓	✓	✓	✓
Structured Data Completeness	✓	✓	✓	✓
Service or Product Schema	✓	✓	✓	✓
FAQPage Schema Present	✓	✓	✓	✓
AI Live-Fetch Speed	✓	✓	✓	✓
Content Visible Without JavaScript	✓	✓	✓	✓
Content & Expertise
Content Recency Signals	✗	✗	✗	✗
Homepage Content Length	✗	✗	✗	✗
Content Structure Readability	✗	✗	✗	✗
Target Customer Stated	✗	✗	✗	✓
Specialization Declared	✗	✓	✓	✓
Language or Population Served	✗	✓	✓	✓
Pricing Transparency	✗	✓	✓	✓
Case Study or Portfolio Presence	✓	✓	✓	✓
Title Tag Not Keyword-Stuffed	✓	✓	✓	✓
Testimonial or Review Content	✓	✓	✓	✓
FAQ Content Present	✓	✓	✓	✓
Author or Person Schema (Expertise)	✓	✓	✓	✓
Social Presence
Facebook Profile Linked	✗	✗	✗	✗
Instagram Profile Linked	✗	✗	✓	✓
LinkedIn Profile Linked	✗	✓	✓	✓
YouTube Channel Linked	✓	✓	✓	✓
Yelp Profile Linked	✓	✓	✓	✓
Trust & Safety
Privacy Policy Availability	✗	✗	✗	✗
Terms of Service Accessibility	✗	✗	✗	✗
Secure HTTPS Connection	✗	✗	✗	✗
User Data Handling Clarity	✗	✗	✗	✓
Third-Party Trust References	✗	✗	✓	✓
Payment Trust Seals	✗	✓	✓	✓
Trustpilot Badge or Reference	✗	✓	✓	✓
Google Trust Badge Shown	✓	✓	✓	✓
Content Security Policy	✓	✓	✓	✓
Secure Resource Loading	✓	✓	✓	✓
Valid SSL Certificate	✓	✓	✓	✓
Security Headers Present	✓	✓	✓	✓
SSL Certificate Not Expiring Soon	✓	✓	✓	✓
Refund and Cancellation Policy Published	✓	✓	✓	✓
Guarantee Statement Present	✓	✓	✓	✓
BBB Reference Present	✓	✓	✓	✓
AI Corroboration
Google Business Profile Links This Domain	✗	✗	✗	✗
Independently Recognized on Yelp	✗	✗	✗	✗
Business Name Consistent Across Sources	✗	✗	✗	✗
LinkedIn Profile Matches the Business	✗	✗	✗	✓
Phone Consistent Across Sources	✗	✗	✓	✓
Address Consistent Across Sources	✗	✓	✓	✓
Independently Recognized by the BBB	✗	✓	✓	✓
License Verified Against State Source	✓	✓	✓	✓
Independently Recognized on Trustpilot	✓	✓	✓	✓
Corroborate the founding date with an offsite source	✓	✓	✓	✓
Displayed Rating Matches Google	✓	✓	✓	✓
Posted Hours Match Google	✓	✓	✓	✓
Score	58%	64%	71%	78%

The Objection: "The Little Things Don't Matter"

The experienced answer-engine skeptic has a fair position. It's mostly reviews and mentions and outside authority, they might say, and all these small on-site factors are rounding error. Let me give you two situations that explain why that is likely wrong.

From the other side, both buildings look identical.

But a thorough review would prove they are not.

You just interviewed someone. You saw the suit, the tie, the nice watch, the fresh haircut, the pressed shirt. So you make your decision and reach across the table to shake their hand, sure you've got them read. Then, as they leave the conference room, you notice they're wearing shorts and no shoes. That's what the full factor list is for.

Or, you ask a strong candidate to come in. They have your address, they sound perfect on the phone, they answer every question right, and they tell you they'll be there in an hour. Three hours later, nobody has shown up. Eventually you work out that you told them Suite B in Building C, and they've been sitting in Building B, Suite C the whole time. You thought you had everything you needed to place them. One character kept them from reaching you.

It doesn't matter how good your ranking strategy is — if the basics are not in place, the engines can't be sure. They are all racing to be the “most trusted” AI so they can't afford to send someone to an abandoned building or to Joe's in the wrong part of town.

Which Categories Carry the Most Weight

After running tens of thousands of tests, we have validated what the search community has said were key factors in ranking in AI. We can definitively say that while you’re laying the foundation, GBP, reputation, and corroboration are the big rocks to work on to move the needle the fastest.

A category’s weight is derived from the individual factors associated with it. All factors are reweighted from new scans each month so we can track changes in AI preferences. We publish the new weights on the 5th day of the following month. The weights in the table are from June 2026.

The Top Three Groups, in Order of Weight

Most of what the SEO and AEO field already teaches lines up with what our data shows.

1. GBP and consistent identifying factors

An answer engine's first challenge is identity resolution. It has to decide that the website, the Google profile, the Yelp page, and the directory listing all belong to the same company. It does that by matching the name, the address, and the phone across those sources. When they all match, the model can fuse everything it knows into one confident record. When they don't, the record fractures. A phone that's off by a digit, a suite number on the site but not on the profile, "& Sons" in one place and "and Sons" in another — any of those will do it. And a fractured record is a business the model can't recommend with any confidence.

Cross-source consistency is where the standard shows its teeth most plainly — whether the same name, phone, address, and rating turn up identical everywhere an engine looks. Three real cases from live domains, names removed:

The wall of phone numbers. On one plumbing company, the scan pulled four different phone numbers scattered across the site, not one of them matching the single number on its Google profile. This is rarely the owner's doing — it is usually the website builder, a booking widget or a call-tracking script seeding phone-formatted strings into pages the owner never sees. But the machine sees all four and has to decide which one is the business. Hand it one number, everywhere, and there is nothing to untangle.

The name that wasn't the name. Picture Willie's Icehouse — a hand-painted sign, a gravel lot, a crowd that's called it Willie's for 20 years. Ask anyone in town and they'll point you straight to it. Now watch a machine try. The sign says Willie's Icehouse. The business license, filed with the state, reads Williams Restaurant & Bar. The Google profile — where the reviews actually live — shows Willie's Ice House & Grill. And the local news station that ran a segment on the place calls it Willie's Bar and Grill. Four names, one building. A person reconciles all four without a second thought: the same beloved dive whether the paperwork says Williams or the sign says Willie's, whether it's "Icehouse" or "Ice House" or the anchor's "Bar and Grill." A machine can't — it sees four separate businesses, confirms none of them, and recommends the competitor down the road whose name reads the same in every place it appears.

The name that was a page title. On one plumbing company, the name the site handed the engine wasn't a name at all — it was the SEO page title, "Emergency Plumber in Springfield · Drains, Leaks & Water Heaters," while the Google profile, the directories, and the lettering on the truck all read a plain company name. Chasing keywords, the site told the machine a different business than every other source listed.

The rating that didn't add up. On a third, the site displayed 834 reviews; Google showed 1,075. The scan says it plainly: "site shows 834 reviews vs Google 1,075 — show your real Google rating, not an inflated or stale one." A stale review widget was quietly undercutting the one number a prospect trusts most, shown in a single line.

2. Reputation: the reviews

Reputation is the second most heavily weighted category. Only the completeness of the Google profile counts for more. It is also the category businesses score worst on. Most businesses have not built it, and fixing it costs a few minutes a week.

When an answer engine recommends a business, it puts its own credibility at risk. It is telling a real person to go use that business.

Reputation is not one number. The engine weighs four separate parts of it. The first is volume: enough reviews to look like a real, active business, not three reviews from 2021. The second is recency: new reviews need to keep arriving, because a business that stopped collecting reviews looks like a business that stopped operating. The third is response: whether the owner replies to reviews, which makes the owner look engaged and accountable. The fourth is distribution: the engine looks past the average star rating and checks whether the low ratings all name the same complaint.

Response is the easiest of the four to fix, and it is heavily weighted. Yet across the businesses we scanned, only 0.8% reply to their reviews. That 0.8% is the share that reply at all, not the share with good reviews. Replying costs only a few minutes a week.

Reputation is also the one signal a consultant cannot fake or rush. It has to be built. The business asks its happy customers for reviews on a regular basis, and it handles the unhappy customers in public.

There are dozens of ways to increase reviews:

3. Corroboration: outside sources back up the claims

At the beginning of 2026, corroboration was a minor factor. Every month since, at every reweighting, we are seeing it rise in prominence. Anyone can type "licensed and insured" on a homepage. Far fewer businesses link that claim to the state licensing board's verification page. A founding year printed in the footer is only a claim. The same year listed on an independent business registry is a fact the engine can check. Corroboration is the difference between a business making a claim about itself and an outside source confirming that claim.

National Chains Get A Head Start

This one will save you from burning a client's budget in the wrong place. Your local businesses are not going to out-rank the national chains on the generic head terms, the "pizza near me" and the "dentist near me." The chain's brand is corroborated all across the web. The model has seen it again and again, and even a single mediocre franchise location inherits all that strength. We watched it happen right in the data. One national chain scored near the bottom of the factor list on its own site, and it still came back as the top AI recommendation, on the brand alone.

Local businesses win on the specific — the neighborhood, the niche service, the odd-hours or odd-need question the chain doesn't serve. A model routes a specific request to a specialist, because specificity is exactly the thing it can match and a chain's generic profile can't. "Emergency Sunday plumber in the Heights who works on tankless systems" is a query no franchise is set up to win. And every well-built local business can.

This is why the factor list matters more for a local business than for a chain, not less. The chain is coasting on corroboration it piled up over decades. The local business has to build its case on purpose, deliberately, and the factor list is that case, item by item.

The Work List

By now it seems like everyone has an AI visibility tool; there's one in every browser.

But what good is knowing Joe's Auto is behind Main Street Repair if you don't know why?

That is where the factor list earns its keep. You don't just see the percentage Joe is behind; you see the exact factors Main Street has in place that Joe doesn't yet. That gap is filled with things that can be done immediately by the agency to show their client their activities while they are waiting on the longer-tail content strategies to kick in.

I can't promise that if Joe aligns his on-site and off-site factors to match his competitors he will pass them. Every case is different, and Joe may never catch Main Street. But our data is clear on one thing: the businesses that focus on aligning these foundational factors get AI's attention a lot faster than the ones that don't.

Here is what two months of that work looks like. As Joe's score climbs toward Main Street's, AI names his shop more often. As the aligned factors increase the mentions generally follow. Having the factor checklist to follow puts clients at ease because they see weekly improvement in alignment while they are waiting for mentions to increase.

The Gap Analysis

A green mark means the factor is in place. A red mark means it is missing. This is a typical situation in which a client is only a few factors away from their competitors. In this example the client comes in with three of the nine factors in place.

Prominent Ranking Factor	Client	Top Ranked A	Top Ranked B	Top Ranked C
Business Name In Title	✗	✗	✗	✗
Business Category Clarity	✗	✗	✗	✗
Years In Business Signal	✗	✗	✗	✓
Location Validation Strength	✗	✗	✓	✓
Team & Role Transparency	✗	✓	✓	✓
Domain Age Verification	✗	✓	✓	✓
Business Email Configured	✓	✓	✓	✓
About Page Present	✓	✓	✓	✓
Legal Entity Type Disclosed	✓	✓	✓	✓
Score	58%	64%	71%	78%

Often it is just a few factors that separate a business that doesn't rank from the next one that does.

The gaps are easy to see. Each of the three top-ranked businesses carries more of these factors than the client does. Every red mark in the client column with a green mark beside it is a specific thing to fix.

The Task List

Every factor has a known fix, and every fix gets logged as it is done, so the client can see the work happening. Click any factor to open its full detail in the Appendix.

Prominent Ranking Factor	Client	Task List
Business Name In Title	Fixed	Added the business name to the homepage title tag.Mac · Mon
Business Category Clarity	Fixed	Stated the service category in the title and main heading.Mac · Mon
Years In Business Signal	Fixed	Added a tenure line to the footer.Sue · Tue
Location Validation Strength	Fixed	Added address, hours, map, and LocalBusiness schema to the Contact page.Mac · Wed
Team & Role Transparency	Fixed	Added a Team page naming the owner and roles.Sue · Thu
Domain Age Verification	✗	Can’t be changed. A young domain will always fail this one, and that is just life.
Business Email Configured	✓	Already in place.
About Page Present	✓	Already in place.
Legal Entity Type Disclosed	✓	Already in place.
Score	89%

The Manual Process

Scoring one business against all 141 factors by hand takes hours, and an agency doing this for many clients cannot keep doing it that way. AI can take on part of the work. You can have an AI model write code that reads a website and checks the factors found on the page itself, such as the title tags, the headings, the contact details, and the schema markup. That covers a meaningful share of the list, and you can build it with the AI tools you already have.

The rest of the work is harder, and doing it well is a real engineering project. Two things make it hard. First, a business's information does not stay the same over time, so its website has to be scanned again on a regular schedule to catch what has changed since the last scan. Second, an AI model working on its own is not reliable at this task. It will report that a factor is present when it is actually missing, and report a factor as missing when it is actually present. On a single business you could catch those mistakes by hand, but across hundreds of businesses you cannot.

Everything in this book you can verify for your client by hand. In fact we did check all 250 of the starting factors by hand several times to make sure the programs we wrote were accurate. Does the name match, letter for letter, on the site, on the Google profile, and across the directories? Is the phone a real tap-to-call link on a phone, or just a picture of a number? Is the schema there, and the right kind? Do popups shove themselves in front of the content before a reader gets to it? Are there negative reviews sitting months old with no reply? With a good programmer and AI you can check just about everything, and it only took us a few days each time.

You can lean on AI to move through it faster. But keep one thing in mind while you do. An AI will sometimes tell you what it thinks you want to hear instead of what’s actually on the page, and it says both with the same confidence. So anything you don’t check with your own eyes, treat as a maybe. Count on some of the data you don’t personally verify being dead wrong.

Then, once you’ve scored the client and a dozen competitors, the real job starts. You fix the site, build the offsite proof, get the reviews answered, and a week later you do it all again, the same way, so you can see what your work changed.

This is why a purpose-built system does the job better than a person, and better than a person using AI alone. If you're serious about building a system to check these factors yourself, it is best to build a separate evaluator for each factor and then store the result every time you run it.

The Easy Way To Track Everything

I’m hoping you can see how the ‘standard’ foundational factors would help you, as an agency or consultant, rank your clients. It was not my intention to turn this book into a sales brochure for my company. I could have done that with a brochure, a webpage, or an article. But almost a year ago I struggled to check these factors, manually and even programmatically, and it is a real challenge.

I realize that now that I’ve published the full list, anyone could upload it into a database and use it. I expect competitors to build their own scanning radar to analyze the prominent factors too. But I don’t think anyone will ever build a better mousetrap to rank local businesses than what we built, and I am proud of it.

Here’s what viizable does for you as an agency, solopreneur, or consultant.

First, like any AI visibility tool, you can research mentions and citations for any domain. Since there are hundreds of variations, we’ve set it up so with a click you can move just the ones you want to include in the weekly trend reports.

In the Visibility tab we track prompts, stats and trends to show progress for the client week to week. Once a prompt to track is added, it will be included in a weekly rescan for the agency and client progress reports and trend charts.

Does this look familiar? Just enter a domain and we automatically check the 141 factors for you and drop them in the workbench—and we select, out of the 50 nearest, the 5 AI-recommended businesses that rank higher than you. But all the hard work is done—you just need to work on aligning the factors so your client can catch their competitors. Just click down to see the detail.

When you click down on a category, you get the factors in that category. When you click a factor, you get the details, and the details show you the data the evaluator found that failed the test. And we tell you exactly how to fix this factor. Plus you click the task button and assign the fix right to the client’s tech guy, or your tech contractor, or even back to the client for research. Working the factors is as easy as pushing buttons all day.

We drilled down on a category, then on the failed factor to get the detail, and from the detail we click the Task button and here we are—everything is completely filled, just assign it to someone and put in a due date.

As you work or your team works, you can easily keep track of everyone’s assignments, delivery dates, and more. And all the factors you fix roll right up to the client and agency reports, showing each week’s progress.

We don’t trust AI to weight or scan the factors, but we do use it to find helpful insights that will allow you to identify the quick wins and impress your client. They are different, and specifically drafted based on each domain’s unique situation and data.

Our analytics are second to none. We provide dozens of reports for both the agency and the client, so there’s always that perfect report for the special situation. Our favorites are the trend reports that show, as you fix the foundational factors, the weekly scan picks them up, and the mentions follow a few weeks later. Lots of rising trend lines in the viizable system.

Since marketing is the backbone of any business—we built our agency clients a “marketing machine.” It uses our best technology to scan prospects’ domains and their competitors and compile the perfect comparisons, gap analyses, and proposals to win new clients. Bright, colorful, impactful—all branded with your agency logo and theme. Dozens of templates from social media, postcards, brochures, emails, PDFs…all showing your prospect why they need you. We’re still building, so bring your ideas and let’s make marketing fun again.

Everything is ‘push a button’ simple.

Final Thoughts About Ranking

Not sure the foundational factors matter? Still not sure what the key to ranking a business is? Whether you use any of the information in this book or not, these are the big rocks that can help you move your client into the rankings.

Focus on consistency. The business's name, phone, and owner have to match everywhere on the web — the site, the Google profile, the LinkedIn page, any old listing a past agency left behind. When they don't match, AI sees more than one "Joe" and can't tell which is the real business.

Focus on content. Right now Joe's site says "great service, honest prices," the same words as the shop next door. What Joe actually does is install glass-pack mufflers, carry one specific national tire brand, and work on old and rare cars other shops turn away — and none of that is on his site yet. Putting it there is the work, because specifics are what AI picks up.

Focus on community. Reviews are a big part of what AI weighs. Joe needs recent ones across more than one platform, like Google and Yelp, and the owner answering them — steady, answered reviews spread around do more than a pile of old ones on a single site.

Focus on corroboration. A claim on Joe's own site is stronger when something outside it agrees. If he says he's ASE-certified, the certifying body should show it; if he sponsors the local team or gets written up in the paper, that should be easy to find. A claim backed up somewhere else is worth far more than one that only lives on Joe's site. If possible, get publishers who write about Joe to link back to the mention on Joe's page.

I think if you got this far in the book, you know that tracking client domains and their competitors week after week across all the factors is nearly impossible. Even with AI you'd need to build at least one evaluator per factor and then find that data, hook up the collectors, and then build the reporting structure. But you don't have to because we did it for you. And because no one else built tools to attract or track AI, we made those as well.

Besides viizable.com, we built LLMSMaker.com, which is an extended LLMS that the AI engines say they love. And since no one had tracking to show us whether AI was even visiting our client sites, we built AISiteStats.com, a free tool that you snap into any site, and then you start to see the AI-based crawlers and bots and the pages they care about. With these three tools, you can leave the debates behind and start doing the things that can cement you as an expert in this new field.

viizable

Need to see which factors are present for your client's domain? Need to see how the foundational factors for the 5 competitors around your client stack up? Just drop in your domain. In less than 5 minutes you are looking at a table with your client and five competitors, the delta marked for you. Every factor scanned and scored, in less time than it takes to cook a small bag of popcorn. Need to know what to do to fix each factor? The full resolution is right there. Need to assign the ‘fix the website title’ to a tech at the client, or to a contractor? That's built in. Need to track the factors you fixed this week? It does that. Need to see a chart showing the factors getting fixed each week and the client's mentions increasing as well? It does that. Want to scan a prospect's domain and send them a professional proposal that alerts them to the factors that need aligning, and how close they are to their competitors? It does that.

If you use viizable, the daunting task of checking the alignment of websites is as simple as pushing a button. Tracking the work and proving to your client that you're hustling is as simple as pushing a button. And marketing to prospects using dozens of branded templates, with deep analytical data yet simple to understand, is as easy as pushing a button.

Most AI-visibility tools stop at mentions and citations. They tell you where a client got named, which is worth knowing.

I spent almost a year analyzing data to determine what the top AI businesses had in place, and then built the tool to manage not only the content projects, but the rest of the stack. Grok just compared us to the leaders in the AEO/GEO/SEO transitioning tool providers, and named us the Top Tool For Ranking Local Businesses In AI.

LLMS Maker

There's a small file that is stirring up a lot of controversy that could turn out to be extremely valuable. The file is called llms.txt. It's a plain-text file you place at the root of your site. Jeremy Howard of Answer.AI proposed it in September 2024, and it's specified over at llmstxt.org. The idea is simple. You hand an AI a clean, plain-language guide to your most important content, so it doesn't have to piece your business back together from scattered HTML.

It got sold as a silver bullet. Publish llms.txt and the answer engines will find you. Then the skeptics showed up with data, and on the data, the skeptics are right.

Here's what they have, and it's fair. Google has said publicly that it doesn't use llms.txt and has no plans to. Independent studies have gone looking for a link between publishing one and getting cited by an AI, and they came up empty. And when you actually watch the crawl logs, which we do, the major answer engines almost never fetch the file at all. So on the plain question — does publishing a standard llms.txt get a local business recommended — the honest answer today is no.

But there's a wrinkle the skeptics undersell. llms.txt is real, and it is being adopted, just not for search. Where it has caught on is software. Technology companies publish one so that coding and documentation agents can read their docs cleanly, and for that job it works fine. That's the reason the local-business version fell flat. It was the right file aimed at the wrong job.

And that's the clue to the wrong question everyone was asking. The studies all measured the plain file — a name, a description, a list of links. Of course that carries no weight for a recommendation. It's just a business talking about itself in a tidier format, and self-description is the exact thing the models discount.

So we ran a different test.

We took one business, a restaurant, and built its llms.txt two different ways. The first was the standard version everyone publishes. The second was that same file rebuilt as a verifiable evidence manifest, with every claim paired to the independent source that confirms it. The license linked to the state board. The reviews pointed at the platforms that host them. The founding date matched to the public registry. The credentials tied to the bodies that issued them. Then we handed both versions to all five engines and asked each one what the file did for its confidence in the business.

The other four said the same thing about the corroborated file, in their own words.

The plain version got exactly what the skeptics predicted, a shrug that told the engine nothing about whether the business was real or good. The evidence-manifest version flipped every single one of them. The difference was the LLMS Maker format and data.

What that means is narrower than it sounds, and it's worth saying plainly. Five engines rewarding a corroborated file could mean the answer engines are ready to lean on a standard like this. But they're independent companies, each one racing to be the one people trust most, and whether any of them formally adopts a format is their call, not ours. What we can control is the input. And that's the whole of what we build: files and formats, generated from a business's own verifiable facts, that raise its corroborative value for the engines that do look. All of it without putting the business's existing organic search at risk. You're adding proof a machine can check, and you aren't gambling your Google ranking to do it.

So let me be exact about what makes the difference, because it isn't the file itself — it's how the file is built. Most tools that make an llms.txt hand you a flat list. Here's our name, here's what we do, here are a few links. Every line on it is something the business said about itself. It's a résumé, and you already know what a résumé is worth, because you've read a hundred of them.

The corroborated version is built the other way. Every claim gets broken into four parts: the value, the source it came from, whether it's been verified, and the link that proves it. "Founded 2009" stops being a sentence on a page and becomes a value — 2009 — tied to the Secretary of State's formation record, marked verified, with the government link sitting right there for the machine to confirm in a single hop. Do that across a whole business and you haven't written a summary. You've assembled something closer to a security-clearance application, where every line is checked against somebody other than you.

Set the two files side by side. The flat one says "licensed and insured," and the machine has nothing but your word for it. The built one says "license number C-48213, verified," with a link straight to the state board that confirms it, and the machine clicks once and knows. Same business, same spot on the same site. One is a résumé and one is a clearance.

A flat llms.txt is a résumé. A corroborated one is a security clearance.

AI Site Stats

ChatGPT, Claude, Gemini, Perplexity and Grok read pages around the clock. AI Site Stats shows you which engine read which page, and how often. It installs as one line of code.

What earns it a place in a client meeting is the split it draws. In a raw server log every bot hit looks alike. AI Site Stats sorts them into two piles.

The first pile is the index crawlers. GPTBot, ClaudeBot, Google-Extended, PerplexityBot. Those are engines reading your pages to build up what they know about a business, on their own schedule, in the background. That work is how a business ends up inside the model at all, and it tells you nothing about today.

The second pile is the on-demand fetchers. ChatGPT-User, Perplexity-User, Claude-User. Those only fire when a person is sitting in a conversation at that moment and the engine goes out to read the page so it can answer them. That is a customer, mid-question, with your client's page open in front of a machine that is deciding what to say about them.

It's free, and it only takes 5 minutes to sign up, set up, and start tracking.

About the Author

J. Brent Tuttle has spent his career turning marketing into something you can measure. He grew up in El Campo, Texas, left college to run his father's insurance and real estate business, and then opened his own agency in Houston. Working from W. Edwards Deming's method — try something, measure the result, improve it, and try again — he grew that agency fast enough to be named the Fastest Growing Insurance Company in Houston.

In the early days of the web, he served as Director of Internet Marketing for a venture-backed startup, competing to rank its products back when Google was still running third behind Yahoo and AltaVista. Decades later he watched the same wide-open moment arrive with AI — and got to work.

He built the scanning engine and the study of nearly 30,000 businesses behind this book, along with the tools that put its findings to use: AIBusinessScore.com, where any owner can see where they stand, and viizable, the workbench built for the agencies and consultants doing the work. He wrote this book to hand that research to the people who can use it, while the window is still open.

You can find his work at aibusinessscore.com and viizable.com.

This appendix walks the factor list, one category at a time. A few things to hold in mind as you read.

It's organized by weight, not by tradition. The categories and the factors are ordered by weight, which is our monthly read of how much each one actually counts. That's why the Google profile and reputation lead the list, and social trails, no matter how much attention the old SEO world paid to each. Every factor carries a weight from that monthly reweighting. Treat it as the current read, not a permanent constant, because the models shift and so do we.

Each category has the same shape.

• Why the category matters and where it ranks.
• What the engines themselves said about it, in their own words.
• How the average business actually scored, factor by factor, with the pass rate from the study.
• A close look at the factors that move the needle — what's available to the AI, why it matters, and how to fix it.
• And last, the rest of the category at a glance.

Every factor is entered in full in Appendix B.

#	Category	What it covers	Factors
1	Local SEO & Google Business Profile Signals	Whether the Google Business Profile exists, is complete, and agrees with the site. The listing itself, its category, core fields, weekly hours, photo coverage, Google Posts, service area, map embed, and the name, address, and phone written on the site.	11
2	Reputation Signals	The review record read as a shape rather than a score. Volume, freshness, cadence, sentiment balance, rating distribution, owner responses, repeated complaints, what the business is known for, and any award stated where it can be read.	11
3	Business Identity Signals	The facts that establish a real, identifiable business. Name in the title, stated category, years in business, validated location, domain age, an email configured on the domain, an About page, named people and roles, and the legal entity.	9
4	AI Visibility	Whether the engines can reach the page and read it. Crawler permissions across the index bots and the on-demand fetchers, indexing and robots.txt rules, live-fetch speed, whether the content survives with JavaScript off, an llms.txt file, and the schema carrying the facts.	22
5	AI Corroboration	Whether independent sources confirm what the site claims. Name, phone, address, displayed rating, and posted hours matched against the Google profile. Recognition on Yelp, LinkedIn, the BBB, and Trustpilot. The license checked against the state source and the founding date against an offsite record.	12
6	Content Quality & Expertise Signals	What the site states about who it serves and what it knows. Target customer, declared specialization, language or population served, pricing transparency, case studies, testimonials, FAQ content, content recency and length, readability, and author schema.	12
7	Trust & Safety Signals	The published terms and the security underneath them. Privacy policy, terms of service, data handling, refund and cancellation policy, guarantee statement, third-party and payment seals, BBB and Trustpilot references, and the SSL certificate, security headers, and content-security policy.	16
8	Contact Prominence	Whether a customer can reach the business from the page. Email address, contact page, accessible form, clickable phone tags, visible physical address, posted operating hours, and calls to action.	7
9	SEO — Technical & On-Page	The familiar fundamentals, weighed against what the recommended businesses actually carry. Load speed and server response, title, meta, H1, canonical, and viewport tags, sitemap and robots.txt, redirects and errors, internal links, image optimization and alt text, responsive layout, page weight, and schema depth.	26
10	UX & Conversion Signals	What sits above the fold, whether booking or availability is offered, whether popups obstruct the page, and the range of calls to action.	4
11	Service Trust & Eligibility Signals	The vetting markers a service business carries. Services described, licensing and insurance stated on the site, the owner identified, contact options available, and local profile completeness. Applies mainly to the trades.	6
12	Social Presence Signals	The lightest-weighted category. Whether the site links its Facebook, Instagram, LinkedIn, YouTube, and Yelp profiles.	5
	Total		141

Local SEO & Google Business Profile Signals

Your Google Business Profile isn't a marketing nicety anymore. To AI, it's your birth certificate.

Why this category matters, and where it ranks

When someone asks an AI for "a good plumber near me," the machine has to do two things before it can answer: figure out which businesses are real and where they are, and then figure out which ones are near the person asking. The Google Business Profile (GBP) is the structured record it reaches for first. It's the canonical, machine-readable version of a business — category, hours, location, phone, photos, reviews — and the answer engines lean on it harder than on anything written on the site itself.

Here's the hard truth this category exposes: a business with no Google Business Profile is, to AI, very close to a business that doesn't exist. It can't be placed on a map, can't be matched to "near me," can't be recommended with confidence. And yet the average business passes only 34% of the checks here — often because the profile exists but is half-built, points at the wrong website, or says something different from the site it's supposed to match.

AI Opinion Of The Month

AI engine	Importance rating	In its own words
Perplexity	10	Strongly affects local discovery and entity matching for location-based intent.
GPT-4o (OpenAI)	8	Key for ensuring local discoverability and relevance.
Grok (xAI)	7	Directly impacts local discoverability and matching to user queries.
Gemini (Google)	7	Indicates relevance and discoverability for local queries.
Claude (Anthropic)	6	Important for local discovery, though it leans harder on direct trust signals.

Higher = weighs heavier in the recommendation decision. AI engines weigh in each month so we can see how closely they concur with our weighting formula. The formula is based on prominence factors present in actual top-ranked businesses since the last weighting period. While collectively they may agree, they usually vary wildly and thus we can't use them in weighting the factors.

What the average business actually scored

Factor	Weight	% pass
Active Google Business Profile	2.0	48.9%
NAP Elements Present On Site	2.0	46.3%
GBP Listing Present (domain-matched)	1.9	23.0%
GBP Category Aligns With Site Content	1.7	27.2%
Photo Coverage Strength	1.5	24.6%
GBP Core Profile Fields Complete	1.3	46.6%
GBP Core Listing Fields Present	1.3	46.5%
Contact Channel Readiness	1.2	43.0%
Map Embed Present	1.0	29.6%
Service Area Specificity	1.0	5.9%

Full entries for every factor are in Appendix B.

The factors that move the needle

The rest of this category, at a glance

Factor	Weight	% pass	What it checks (full entry in Appendix B)
Photo Coverage Strength	1.5	24.6%	Fresh, sufficient photo coverage on the profile — an engagement and credibility signal.
GBP Core Profile Fields Complete	1.3	46.6%	Description, phone, address, and category all populated.
GBP Core Listing Fields Present	1.3	46.5%	Hours, phone, address, category, and at least one photo.
Contact Channel Readiness	1.2	43.0%	Multiple ways to reach you — phone, booking link, messaging.
Map Embed Present	1.0	29.6%	An embedded map that reinforces your location to AI and visitors.

Note: While we query AI engines for the weight that they may put on certain factors or categories of factors, we reweight based on actual prominence data from our scanning and evaluations of top recommended businesses. Figures as of June 2026.

Reputation Signals

What customers say about you — and whether AI can find it said in more than one place.

Why this category matters, and where it ranks

Prominence is very high in this category, and although we don't use the AI engines to weigh the factors, we do want to see if they corroborate our data. When we asked GPT, Gemini, Grok, Perplexity, and Claude, in their own words, what moves their decision to recommend a local business, reputation came back first or second on every single engine.

Here's the catch, and it's the whole point of this chapter: this is the category local businesses fail the hardest. Across the businesses we scanned, the average one passes only about 28% of the reputation checks. The signal AI trusts most is the signal most businesses are weakest on. That's not bad news — that's the opening.

AI Opinion Of The Month

AI engine	Importance rating	In its own words
Claude (Anthropic)	14	Strongest signal of actual customer satisfaction.
Grok (xAI)	13	Strong evidence of reliability and satisfaction.
Gemini (Google)	12	Directly reflects satisfaction and builds trust.
Perplexity	12	Volume, freshness, sentiment = core trust inputs.
GPT-4o (OpenAI)	10	Heavily influences consumer trust and ranking.

AI engines weigh in each month so we can see how closely they concur with our weighting formula. The formula is based on prominence factors present in actual top-ranked businesses since the last weighting period. While collectively they may agree, they usually vary wildly and thus we can't use them in weighting the factors.

What the average business actually scored

Below is every reputation factor we measure, sorted by how heavily it weighs, with the share of real businesses that pass it. Read the right-hand column as a map of where the money is: the lower the pass rate on a high-weight factor, the bigger the opportunity sitting untouched.

Factor	Weight	% that pass
Review Volume & Recency	2.0	40.6%
Review Volume Strength	1.8	44.0%
Negative-Review Pattern	1.8	36.8%
Review Freshness	1.7	40.7%
Review Sentiment Balance	1.6	45.1%
Rating Distribution Health	1.5	41.3%
Multi-Platform Review Presence	1.5	14.0%
Directory Citation Presence	1.5	6.0%
Responsiveness to Reviews	1.4	0.8%
Review Distribution Across Platforms	1.3	4.8%
GBP Photos & Reviews Present	1.2	34.3%

Every factor in this category — and the other eleven categories — is listed in full in Appendix B.

The factors that move the needle

The corroboration pair — where Reputation overlaps the new AI rules

Two reputation factors aren't really about reviews at all. They're about whether your reputation exists in more than one place — and they're the two almost nobody passes.

The rest of this category, at a glance

Factor	Weight	% pass	What it checks (full entry in Appendix B)
Review Sentiment Balance	1.6	45.1%	Whether tone skews positive, even when the star average looks fine.
Rating Distribution Health	1.5	41.3%	At least 60% five-star, no more than 10% one-star — shape beats the raw average.
Review Distribution Across Platforms	1.3	4.8%	Volume plus breadth across listings; concentrated-on-one-platform reads as narrow.
GBP Photos & Reviews Present	1.2	34.3%	Profile carries both customer photos and reviews — harder-to-fake authenticity.

Note: While we query AI engines for the weight that they may put on certain factors or categories of factors, we reweight based on actual prominence data from our scanning and evaluations of top recommended businesses. Figures as of June 2026.

Business Identity Signals

Being recognized as a legitimate business is a key foundational factor.

Why this category matters, and where it ranks

This is the front door. Every other category in this book assumes one thing: that the AI already accepts you exist and are who you say you are. Identity is where that gets settled. It's the third-highest-weighted category across the engines we test, and it's the one that gates everything else — a brilliant review profile does you no good if the AI can't confirm there's a real, locatable, named business attached to it.

Here's the gut-punch: the average local business passes only 35% of the identity checks. Not because owners are hiding — because the basics live in their heads, on a business card, on their Google profile, anywhere except in plain, crawlable text on their own website. AI can't read your mind or your intentions. It reads your page. If the page doesn't say it, to the machine it isn't true.

AI Opinion Of The Month

We put the question to every engine on a rolling basis, and identity came back as a baseline requirement on all of them — the thing that has to be true before anything else counts. Below is the current read.

AI engine	Importance rating	In its own words
Perplexity	12	AI needs to know the business is a real entity before any recommendation is credible.
Grok (xAI)	10	Confirms the business is legitimate and consistently represented.
GPT-4o (OpenAI)	10	Crucial for verifying legitimacy and consistency across platforms.
Claude (Anthropic)	10	Essential foundation for verifying a business exists and is legitimate.
Gemini (Google)	8	A baseline requirement for any recommendation.

AI engines weigh in each month so we can see how closely they concur with our weighting formula. The formula is based on prominence factors present in actual top-ranked businesses since the last weighting period. While collectively they may agree, they usually vary wildly and thus we can't use them in weighting the factors.

What the average business actually scored

Every identity factor we measure, sorted by weight, with the share of real businesses that pass. The low numbers on high-weight rows are the opportunities.

Factor	Weight	% pass
Phone Number Present On Site	1.6	49.5%
Business Email Configured (domain MX)	1.6	49.1%
Title Tag Present	1.5	50.4%
Location Validation Strength	1.5	44.7%
Domain Age Verification	1.5	2.7%
Business Credential Summary	1.4	25.2%
About Page Present	1.4	61.6%
Physical Address Present On Site	1.3	47.0%
Business Category Clarity	1.3	16.4%
Contact Page Present	1.3	35.5%
Legal Entity Type Disclosed	1.3	33.9%
Content Section Headings	1.2	44.1%
Homepage Content Length	1.2	42.7%
Local or Family Ownership Signal	1.2	20.4%
Operating Hours On Site	1.1	44.6%
Named Personnel Present	1.0	49.3%
Years In Business Signal	0.8	4.8%
Insurance Signal On Site	0.5	27.6%
Licensing Signal On Site	0.5	15.4%

Every factor is listed in full in Appendix B.

The factors that move the needle

The rest of this category, at a glance

Factor	Weight	% pass	What it checks (full entry in Appendix B)
Location Validation Strength	1.5	44.7%	Structured, consistent physical-location signals that confirm where you operate.
About Page Present	1.4	61.6%	A reachable About page — where AI and humans learn who you are.
Physical Address Present	1.3	47.0%	A street address in readable page text to anchor local relevance.
Contact Page Present	1.3	35.5%	A contact page at a standard path that actually resolves.
Legal Entity Type Disclosed	1.3	33.9%	LLC / Inc. / Corp. markers that signal a properly formed business.
Content Section Headings	1.2	44.1%	Enough H2/H3 structure for AI to segment and extract your services.
Homepage Content Length	1.2	42.7%	At least ~300 words so engines have substance to work with.
Local or Family Ownership Signal	1.2	20.4%	Family-/locally-/veteran-owned cues for "support local" intent.
Operating Hours On Site	1.1	44.6%	Published hours so AI can answer "open now" queries.
Named Personnel Present	1.0	49.3%	Real human names — owner, staff, team — versus a faceless shell.
Years In Business Signal	0.8	4.8%	An explicit tenure/experience claim ("serving since 2004").
Insurance Signal On Site	0.5	27.6%	Insured/bonded language — a trust cue for service trades.
Licensing Signal On Site	0.5	15.4%	License/certification language for regulated categories.

Note: While we query AI engines for the weight that they may put on certain factors or categories of factors, we reweight based on actual prominence data from our scanning and evaluations of top recommended businesses. Figures as of June 2026.

AI Visibility

It doesn't matter how good your business looks to a human if the AI can't read the page at all. This is the category that barely existed two years ago.

Why this category matters, and where it ranks

Every other category assumes the AI can see your website. This one checks whether that's actually true. And the answer, for a surprising number of modern sites, is no — because the slick JavaScript framework that makes the site feel fast in a browser hands an empty shell to the AI crawlers that don't run JavaScript. The business is invisible at the exact layer where the recommendation happens.

This is the youngest category in the index — half its factors didn't exist as ranking concerns two years ago — and it's the fourth-heaviest by weight. It also contains the single highest-weighted factor in the entire book: whether your content is readable without JavaScript. Most sites pass this category overall (the average is 76.5%, second-highest), largely because you'd have to actively block crawlers to fail the access checks. But the few sites that fail the big factors fail catastrophically: they're not ranked low, they're absent.

AI Opinion Of The Month

AI engine	Importance rating	In its own words
Grok (xAI)	12	Fundamental requirement for AI to access and process business information.
Gemini (Google)	10	If AI can't access the info, it can't evaluate or recommend you.
Perplexity	10	If crawlers can't access the site, the business may not be eligible at all.
Claude (Anthropic)	9	If AI can't parse your info, the business effectively doesn't exist for AI.
GPT-4o (OpenAI)	8	Critical for ensuring AI can access and interpret content.

Notice the language — "doesn't exist," "not eligible at all." This is a gate, not a dial. AI engines weigh in each month so we can see how closely they concur with our weighting formula. The formula is based on prominence factors present in actual top-ranked businesses since the last weighting period. While collectively they may agree, they usually vary wildly and thus we can't use them in weighting the factors.

What the average business actually scored

Factor	Weight	% pass
Content Visible Without JavaScript	2.5	83.8%
Server Response Speed (for live fetch)	2.0	88.5%
Crawler Access Allowed	1.8	98.8%
Open Graph Tags Present	1.8	35.9%
Indexing Allowed (no noindex)	1.7	99.6%
LocalBusiness Schema Present	1.4	30.5%
Structured Data Completeness	1.4	22.3%
Service or Product Schema	1.2	3.2%
llms.txt File Present	1.0	4.0%
AI crawler allow-list (8 bots: GPTBot, ClaudeBot, PerplexityBot, Google-Extended, …)	1.0	~99%

The individual AI-crawler permission checks (GPTBot, OAI-SearchBot, ChatGPT-User, Google-Extended, PerplexityBot, Perplexity-User, ClaudeBot, Claude-User/SearchBot) each pass ~99–100% and are collapsed into one row here; they're listed individually in Appendix B.

The factors that move the needle

The access layer

The remaining factors are the AI-crawler permission checks: whether your robots.txt lets in GPTBot, OAI-SearchBot, ChatGPT-User, Google-Extended, PerplexityBot, Perplexity-User, ClaudeBot, Claude-User, and Claude-SearchBot, plus the general "crawler access allowed" and "indexing allowed" gates. Nearly everyone passes these (~99%) because you have to actively block a bot to fail — but when a site does fail, it's usually an accident (a developer's leftover Disallow, a staging noindex that shipped to production) that quietly erases the business from one or more AI engines. Plus the emerging llms.txt file (4% adoption), a forward-looking signal that you're paying attention to AI discoverability.

Factor	Weight	% pass	What it checks (full entry in Appendix B)
Crawler Access Allowed	1.8	98.8%	robots.txt doesn't blanket-block all crawlers.
Indexing Allowed	1.7	99.6%	No noindex directive hiding the homepage from engines.
8 named AI-bot permissions	1.0 ea	~99%	GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and the on-demand fetch agents are each allowed.
llms.txt File Present	1.0	4.0%	An /llms.txt file signalling AI-discoverability awareness.

Note: While we query AI engines for the weight that they may put on certain factors or categories of factors, we reweight based on actual prominence data from our scanning and evaluations of top recommended businesses. Figures as of June 2026.

AI Corroboration

Anyone can claim anything on their own website. This category checks whether the rest of the internet agrees.

Why this category matters, and where it ranks

Here's the shift that breaks the old SEO playbook. For twenty years, optimization was mostly about what you put on your own site. But an AI that's about to recommend you to a real person has a problem the old search engines didn't worry about as much: it can hallucinate, and it can be lied to. In the businesses that get recommended, the claims are corroborated. Does the claim of a medical degree resolve to the issuing school or a professional association? Does the address on three directories agree? Does the license resolve to a state board record? On the businesses that rank, the answer is yes.

Corroboration is a high-impact category that's easy to fix, where the average business actually scores relatively low at 48% pass.

AI Opinion Of The Month

AI engine	Importance rating	In its own words
Gemini (Google)	23 *	The ultimate validation — the most critical factor for a confident recommendation.
Claude (Anthropic)	12	Independent corroboration is the ultimate validation that claims are accurate.
Grok (xAI)	11	Validates claims across independent sources, preventing misinformation.
Perplexity	10	Cross-source consistency reduces the risk of hallucinated or untrustworthy claims.
GPT-4o (OpenAI)	5	Ensures claims are verified by independent sources, enhancing credibility.

* Gemini rated this category far above the normal range — an unusually emphatic signal. Even setting that outlier aside, the remaining engines place corroboration in the top tier. AI engines weigh in each month so we can see how closely they concur with our weighting formula. The formula is based on prominence factors present in actual top-ranked businesses since the last weighting period. While collectively they may agree, they usually vary wildly and thus we can't use them in weighting the factors.

What the average business actually scored

Factor	Weight	% pass
Phone Consistent Across Sources	1.8	38.2%
Address Consistent Across Sources	1.8	49.7%
GBP Links This Domain	1.5	77.7%
Business Name Consistent Across Sources	1.5	71.3%
Directory Citations Match the Profile	1.5	45.2%
License Verified Against State Source	1.5	6.4%

Full entries in Appendix B.

The factors that move the needle

The rest of this category, at a glance

Factor	Weight	% pass	What it checks (full entry in Appendix B)
GBP Links This Domain	1.5	77.7%	The website on your Google profile is this exact domain, so its authority flows to you.
Business Name Consistent	1.5	71.3%	Your name matches between site and GBP (tolerating legal suffixes), so AI unifies you into one record.

Note: While we query AI engines for the weight that they may put on certain factors or categories of factors, we reweight based on actual prominence data from our scanning and evaluations of top recommended businesses. Figures as of June 2026.

Content Quality & Expertise Signals

AI recommends the business that can prove it knows what it's doing. Most sites only claim it.

Why this category matters, and where it ranks

When an AI weighs two plumbers, it's not just checking that they exist and have reviews — it's reading their content to judge which one actually knows the work. Specific service descriptions, real expertise language, current information, evidence of completed jobs: these are how a machine distinguishes a genuine specialist from a generic template that says "we do plumbing." It's the fifth-heaviest category by weight, and every engine ranks it highly because content is where competence becomes legible.

The average business passes 41% — but the breakdown is brutal and full of opportunity. Sites pass the easy structural checks (a non-stuffed title, a stated specialization) at high rates, and fail the prove-it checks almost completely: current content, specific service detail, first-person expertise, pricing, case studies.

AI Opinion Of The Month

AI engine	Importance rating	In its own words
Claude (Anthropic)	11	Indicates whether a business can competently serve user needs.
GPT-4o (OpenAI)	10	Fundamental for establishing authority and relevance.
Gemini (Google)	9	Lets AI understand offerings and match them to user needs.
Perplexity	8	AI favors content that demonstrates authority and real domain knowledge.
Grok (xAI)	7	Demonstrates expertise that helps AI match you to specific needs.

AI engines weigh in each month so we can see how closely they concur with our weighting formula. The formula is based on prominence factors present in actual top-ranked businesses since the last weighting period. While collectively they may agree, they usually vary wildly and thus we can't use them in weighting the factors.

What the average business actually scored

Factor	Weight	% pass
Content Recency Signals	1.8	0.6%
Content Originality	1.6	34.6%
Service Description Specificity	1.5	8.7%
Title Tag Not Keyword-Stuffed	1.5	95.6%
Content Structure Readability	1.4	36.8%
Target Customer Stated	1.3	89.8%
Specialization Declared	1.3	91.1%
External Links Present	1.2	85.2%
Language or Population Served	1.1	46.5%
First-Person Expertise Signals	1.0	9.3%
Copyright Year Currency	1.0	16.6%
Pricing Transparency	1.0	8.4%
Case Study or Portfolio Presence	1.0	13.0%
Testimonial or Review Content	1.0	14.8%
FAQ Content Present	1.0	13.6%
No Known Stock-Photo Hosts	1.0	90.8%

Full entries in Appendix B.

The factors that move the needle

The rest of this category, at a glance

Factor	Weight	% pass	Note
Content Originality	1.6	34.6%	Distinctive, non-templated writing vs boilerplate.
Content Structure Readability	1.4	36.8%	Clear formatting and headers aid comprehension.
Testimonial or Review Content	1.0	14.8%	First-party social proof embedded on the site.
FAQ Content Present	1.0	13.6%	Q&A format maps directly to how people ask AI.
Copyright Year Currency	1.0	16.6%	"© 2019" signals neglect.
Title Tag Not Keyword-Stuffed	1.5	95.6%	Mostly fine — avoid superlative spam in the title.
Target Customer Stated	1.3	89.8%	Mostly fine — say who you serve.
Specialization Declared	1.3	91.1%	Mostly fine — name specific services.
External Links Present	1.2	85.2%	Mostly fine — at least one outbound link to an authority.
Language or Population Served	1.1	46.5%	Languages spoken / communities served route audience-specific queries.
No Known Stock-Photo Hosts	1.0	90.8%	Mostly fine — original photography reads as authentic.

Note: While we query AI engines for the weight that they may put on certain factors or categories of factors, we reweight based on actual prominence data from our scanning and evaluations of top recommended businesses. Figures as of June 2026.

Trust & Safety Signals

Everyone has the padlock now. AI has moved on to whether you're actually safe and transparent to deal with — and that's where almost everyone falls short.

Why this category matters, and where it ranks

A decade ago, "is this site secure?" meant "does it have HTTPS?" That war is over — 99.6% of businesses now have the padlock, and it no longer distinguishes anyone. So AI has moved the goalposts to the next layer: do you publish a privacy policy and terms? Do you say, in plain language, how you handle customer data? Do you send modern security headers? Do you back your work with a guarantee or refund policy, and reference independent trust bodies? This is a mid-weight category, but it's one where the basics are universal and the depth is wide open.

The average business passes only 37% — because they have the SSL certificate everyone has, and almost none of the policy, transparency, and security-header signals that now do the actual distinguishing.

AI Opinion Of The Month

AI engine	Importance rating	In its own words
GPT-4o (OpenAI)	10	Essential for reinforcing reliability and security.
Grok (xAI)	8	Builds confidence through security, transparency, and policy adherence.
Claude (Anthropic)	8	Important trust markers, especially for sensitive services or transactions.
Gemini (Google)	7	Policies and security reinforce reliability and protect users.
Perplexity	5	Reinforces reliability, but usually supporting rather than primary.

AI engines weigh in each month so we can see how closely they concur with our weighting formula. The formula is based on prominence factors present in actual top-ranked businesses since the last weighting period. While collectively they may agree, they usually vary wildly and thus we can't use them in weighting the factors.

What the average business actually scored

Factor	Weight	% pass
Secure Connection Indicators	1.9	99.6%
Privacy Policy Availability	1.7	28.6%
User Data Handling Clarity	1.6	7.5%
Security Headers Present	1.6	20.6%
Award or Recognition Signal	1.6	18.4%
Terms of Service Accessibility	1.5	17.0%
Refund Policy Published	1.5	1.3%
HSTS Header Present	1.5	17.4%
SSL Certificate Present	1.4	99.4%
SSL Not Expiring Soon	1.4	96.8%
Guarantee Statement Present	1.4	2.9%
Policy Update Freshness	1.4	48.1%
BBB Reference Present	1.3	4.8%
Third-Party Trust References	1.3	15.0%
Mixed Content Free	1.0	99.5%
Content Security Policy	1.0	9.1%

Full entries in Appendix B.

The factors that move the needle

The rest of this category, at a glance

Factor	Weight	% pass	Note
Secure Connection Indicators	1.9	99.6%	Modern TLS. You almost certainly pass.
Terms of Service Accessibility	1.5	17.0%	A findable terms page sets expectations and signals accountability.
HSTS Header Present	1.5	17.4%	Forces HTTPS — part of the security-headers fix.
SSL Certificate Present / Not Expiring	1.4	99% / 97%	Don't let the cert lapse.
Policy Update Freshness	1.4	48.1%	Legal pages dated within ~24 months read as maintained.
BBB Reference Present	1.3	4.8%	A recognized third-party trust reference.
Third-Party Trust References	1.3	15.0%	Badges, associations, audits that establish authenticity.
Mixed Content Free	1.0	99.5%	No insecure resources on an HTTPS page.
Content Security Policy	1.0	9.1%	An advanced security signal — part of the headers fix.

Note: While we query AI engines for the weight that they may put on certain factors or categories of factors, we reweight based on actual prominence data from our scanning and evaluations of top recommended businesses. Figures as of June 2026.

Contact Prominence

Being easy to reach is just good business.

Why this category matters, and where it ranks

A recommendation is only useful if it leads somewhere. When AI hands a customer your name, it's implicitly promising that customer can get in touch — call, email, fill out a form, find the door. If your contact details are buried, broken, or missing, you're a dead end, and AI learns to route around dead ends. This category sits in the middle of the pack by weight, but it's a quiet gatekeeper: it's the difference between being named and being useful.

The average business passes about 35% of these checks — usually failing not because they can't be reached, but because the contact path isn't obvious to a machine reading the page.

AI Opinion Of The Month

AI engine	Importance rating	In its own words
Grok (xAI)	9	Reduces friction in the recommendation-to-action path.
Perplexity	9	Helps verify the business is reachable, accountable, and operational.
Claude (Anthropic)	8	Enables verification and user action — trust signal and practical necessity.
GPT-4o (OpenAI)	6	Ensures users and systems can easily reach the business.
Gemini (Google)	6	Essential for user convenience and verifying legitimacy.

AI engines weigh in each month so we can see how closely they concur with our weighting formula. The formula is based on prominence factors present in actual top-ranked businesses since the last weighting period. While collectively they may agree, they usually vary wildly and thus we can't use them in weighting the factors.

What the average business actually scored

Factor	Weight	% pass
Physical Address Visibility	1.8	39.8%
Contact Form Accessibility	1.6	23.9%
Call-to-Action Placement	1.4	45.1%
Clickable Phone Tags	1.3	42.1%
Email Address Present On Site	1.2	26.2%
Multiple Contact Methods Availability	1.1	32.6%

Full entries in Appendix B.

The factors that move the needle

The rest of this category, at a glance

Factor	Weight	% pass	What it checks (full entry in Appendix B)
Email Address Present On Site	1.2	26.2%	A visible email for customers who prefer to write rather than call.
Multiple Contact Methods	1.1	32.6%	Several ways to reach you, signalling readiness for different customer preferences.

Note: While we query AI engines for the weight that they may put on certain factors or categories of factors, we reweight based on actual prominence data from our scanning and evaluations of top recommended businesses. Figures as of June 2026.

SEO — Technical & On-Page

The familiar category, weighed honestly. The fundamentals still gate whether AI can read you — but most of the old obsessions have quietly lost their grip.

Why this category matters, and where it ranks

If you're an SEO or run an agency, this is the chapter you came in trusting. So here's the straight talk: technical SEO still matters, but not for the reasons it used to. The engines rank it mid-pack — Perplexity gives it a 9 for crawlability, but the rest sit around 6. The reason is a shift in why it matters. Speed and clean structure used to be about pleasing Google's algorithm. Now they're about something more basic: whether an AI can fetch your page fast enough and parse it cleanly enough to use it at all.

This is also the biggest category in the index — 25 factors — and the data tells a blunt story. Businesses pass the structural basics (the site loads, it's reachable, the server responds) at very high rates, and fail the quality basics (image optimization, payload weight, Core Web Vitals, rich markup) almost universally. The average is 43%, the highest of the trust-relevant categories — but it hides a split between "everybody passes" and "almost nobody passes."

AI Opinion Of The Month

AI engine	Importance rating	In its own words
Perplexity	9	Helps AI crawl, parse, and understand the site accurately.
Grok (xAI)	6	Ensures the site is crawlable and ranks in supporting searches.
GPT-4o (OpenAI)	6	Essential for visibility and technical performance.
Gemini (Google)	6	Ensures discoverability; structured data aids AI understanding.
Claude (Anthropic)	6	Helps discoverability and extraction, but isn't a primary trust indicator.

Note the phrase that recurs: "not a primary trust indicator." SEO is necessary plumbing, not the thing that earns the recommendation. AI engines weigh in each month so we can see how closely they concur with our weighting formula. The formula is based on prominence factors present in actual top-ranked businesses since the last weighting period. While collectively they may agree, they usually vary wildly and thus we can't use them in weighting the factors.

The factors that move the needle

The rest of this category, at a glance

The remaining factors split cleanly into "almost everyone passes" (the structural basics) and "almost no one passes" (the quality basics). The near-universal fails are where the easy points hide.

Factor	Weight	% pass	Note
Server Response Time	1.9	94.2%	Mostly fine.
Total Page Response Time	1.8	97.4%	Mostly fine.
PageSpeed SEO Score	1.8	83.9%	Most pass the technical-fundamentals audit.
Desktop PageSpeed Performance	1.6	49.2%	Coin-flip — image/script weight decides it.
Page Payload Weight	1.4	0.9%	Near-universal fail — missing gzip/brotli compression. Easy server win.
Internal Link Structure	1.4	44.7%	Needs ≥5 internal links so AI can crawl deeper pages.
Script Efficiency Signals	1.4	28.2%	Too many heavy scripts slow the page.
Meta Description Quality	1.4	29.5%	Present and ~150–160 chars.
Low Redirect Count	1.4	97.1%	Mostly fine.
Responsive Layout Stability	1.3	44.5%	Elements shouldn't shift as the page loads.
Image Alt Text Coverage	1.3	23.6%	≥80% of images need descriptive alt text.
PageSpeed Best-Practices Score	1.3	40.9%	HTTPS, image ratios, no console errors.
Sitemap Availability	1.3	19.5%	An XML sitemap accelerates AI crawl discovery.
Canonical Tag Presence	1.2	42.9%	Declares the authoritative URL.
Robots.txt Accessibility	1.2	49.0%	Present, ideally with a Sitemap directive.
Site Reachable Without Errors	1.2	99.4%	Mostly fine — if you fail this, nothing else matters.
Viewport Meta Tag Present	1.1	95.5%	Mostly fine — mobile scaling.
Healthy Page Size	1.0	4.3%	Near-universal fail — pages too thin or too bloated.
Review Markup with Data	1.0	4.0%	Near-universal fail — aggregateRating with real values. Big uncontested win if you have reviews.

Note: While we query AI engines for the weight that they may put on certain factors or categories of factors, we reweight based on actual prominence data from our scanning and evaluations of top recommended businesses. Figures as of June 2026.

UX & Conversion Signals

The one category most businesses already pass — which is exactly why it's easy to ignore the part you don't.

Why this category matters, and where it ranks

This is the good-news chapter. Of all twelve categories, UX & Conversion is the one the typical business actually passes — the average score here is 81%, the highest in the index. Modern website builders bake in a clear hero section and block the worst popup behavior by default, so most sites clear the bar without trying.

That's also the trap. Because it ranks lower by weight (the engines treat it as a supporting signal, not a trust anchor) and because most sites pass, it's tempting to assume there's nothing to do here. But the two factors most businesses do miss — clear above-the-fold messaging and a real booking/availability cue — are the ones that turn a recommendation into a customer. Don't pour effort here at the expense of Reputation or Corroboration. Just don't leave the cheap wins on the table.

AI Opinion Of The Month

AI engine	Importance rating	In its own words
GPT-4o (OpenAI)	7	Vital for user engagement and reducing friction.
Grok (xAI)	6	Supports seamless experiences that lead to higher conversion.
Perplexity	6	A clear, low-friction site suggests legitimacy.
Gemini (Google)	5	Ensures a good experience after the recommendation — reflects well on the AI.
Claude (Anthropic)	5	Suggests professionalism, though it doesn't directly validate trust.

A supporting signal — real, but lighter than the trust categories. AI engines weigh in each month so we can see how closely they concur with our weighting formula. The formula is based on prominence factors present in actual top-ranked businesses since the last weighting period. While collectively they may agree, they usually vary wildly and thus we can't use them in weighting the factors.

What the average business actually scored

Factor	Weight	% pass
Above-the-Fold Clarity	1.8	92.2%
Call-to-Action Variety	1.8	80.6%
Booking or Availability Signal	1.2	54.1%
No Aggressive Popups	1.0	98.8%

Full entries in Appendix B.

The four factors

Note: While we query AI engines for the weight that they may put on certain factors or categories of factors, we reweight based on actual prominence data from our scanning and evaluations of top recommended businesses. Figures as of June 2026.

Service Trust & Eligibility Signals

The vetting markers a service business needs — licensed, insured, responsive, accountable. Mostly for the trades, and AI knows the difference.

Why this category matters, and where it ranks

This category is named for the signals behind Google's Local Services Ads — the "Google Guaranteed / Screened" world of licensed, insured, background-checked providers. It's the most industry-specific category in the index, and the AI ratings show it: Claude weighs it a 7 for service businesses, while Gemini drops it to a 3, reasoning it's "less universally critical." They're both right. If you're a plumber, electrician, roofer, lawyer, or any licensed trade, this matters a lot. If you're a restaurant or a boutique, AI doesn't check whether your contractor's license is valid.

It's near the bottom of the index by average weight precisely because it doesn't apply to everyone. But for the businesses it does apply to, these are exactly the trust markers that decide a high-stakes "who should I let into my home" recommendation. The average across all businesses scanned is 29.5% — dragged down by service businesses that simply never state the credentials they actually hold.

AI Opinion Of The Month

AI engine	Importance rating	In its own words
Claude (Anthropic)	7	Indicates rigorous verification — strong trust for service businesses.
Grok (xAI)	6	Compliance with verification standards that prioritize trustworthy providers.
GPT-4o (OpenAI)	5	Important for ad placement, less critical for organic recommendations.
Perplexity	5	Matters for businesses that rely on Local Services Ads — narrower in scope.
Gemini (Google)	3	Less universally critical than broader trust signals.

The wide spread (3 to 7) is the signal here: this category's importance depends heavily on industry. AI engines weigh in each month so we can see how closely they concur with our weighting formula. The formula is based on prominence factors present in actual top-ranked businesses since the last weighting period. While collectively they may agree, they usually vary wildly and thus we can't use them in weighting the factors.

What the average business actually scored

Factor	Weight	% pass
Responsiveness Score Indicators	1.8	24.8%
Service Category Eligibility Match	1.6	41.6%
Complaint & Negative-Signal Indicators	1.6	29.2%
Service Area Eligibility Match	1.5	38.9%
Business Owner Identity Signals	1.4	33.6%
GBP Hours Set For All Seven Days	1.3	44.8%
LSA Profile Completeness	1.3	28.4%
Insurance Signal On Site	1.0	23.6%
License Signal On Site	1.0	0.7%

Full entries in Appendix B.

The factors that move the needle

The rest of this category, at a glance

Factor	Weight	% pass	What it checks (full entry in Appendix B)
Service Category Eligibility Match	1.6	41.6%	Your declared services align with structured data and directory signals.
Service Area Eligibility Match	1.5	38.9%	Geographic coverage is precise enough for location-specific queries.
GBP Hours Set For All Seven Days	1.3	44.8%	Hours defined for every day — an availability completeness signal.
LSA Profile Completeness	1.3	28.4%	Bio, services, and business details fully populated across sources.
Insurance Signal On Site	1.0	23.6%	Insured/bonded language AI can read and cite.

Note: While we query AI engines for the weight that they may put on certain factors or categories of factors, we reweight based on actual prominence data from our scanning and evaluations of top recommended businesses. Figures as of June 2026.

Social Presence Signals

The lightest-weighted category in the index — but the profiles you link still quietly corroborate that you're real.

Why this category matters, and where it ranks

Let's be honest about this one: it's the lowest-weighted category in the entire index. Every engine rated it a 4 or 5, and they agree on why — social presence is supporting evidence, not a trust anchor. A linked Facebook page won't get a bad business recommended, and the absence of a TikTok won't sink a good one. So don't pour your energy here.

But don't ignore it either, because the profiles you link do one useful thing: they're another set of independent places where your business shows up consistently, which feeds the corroboration signal that AI does weigh heavily. Think of social links less as "marketing" and more as additional anchors confirming you exist. The average business passes only 34% here, mostly by linking one or two profiles and stopping.

AI Opinion Of The Month

AI engine	Importance rating	In its own words
Grok (xAI)	5	Community engagement and social proof beyond reviews.
GPT-4o (OpenAI)	5	Additional social proof and brand legitimacy.
Gemini (Google)	4	Less critical than direct performance or reputation.
Claude (Anthropic)	4	Adds legitimacy, but less critical than direct verification and reviews.
Perplexity	4	Useful supporting evidence; usually less decisive than reputation and corroboration.

Striking agreement at the bottom of the scale. AI engines weigh in each month so we can see how closely they concur with our weighting formula. The formula is based on prominence factors present in actual top-ranked businesses since the last weighting period. While collectively they may agree, they usually vary wildly and thus we can't use them in weighting the factors.

What the average business actually scored

Factor	Weight	% pass
Social Profile Count	1.4	18.6%
Facebook Profile Linked	0.5	72.6%
Instagram Profile Linked	0.5	50.0%
YouTube Channel Linked	0.5	24.8%
Yelp or BBB Profile Linked	0.5	20.0%
LinkedIn Profile Linked	0.5	16.1%

Full entries in Appendix B.

The factor that moves the needle

The individual platform links, at a glance

Each of these is worth only 0.5 — together they roll up into the Social Profile Count above. Link the ones that fit your business; skip the ones that don't.

Factor	Weight	% pass	Note
Facebook Profile Linked	0.5	72.6%	The most common — most local businesses already have this.
Instagram Profile Linked	0.5	50.0%	Strong for visual trades (food, beauty, contracting).
YouTube Channel Linked	0.5	24.8%	Worth it only if you actually publish video.
Yelp or BBB Profile Linked	0.5	20.0%	Doubles as a reputation/corroboration anchor — higher value than its weight suggests.
LinkedIn Profile Linked	0.5	16.1%	Best for B2B and professional services; ties to owner-identity corroboration.

Note: While we query AI engines for the weight that they may put on certain factors or categories of factors, we reweight based on actual prominence data from our scanning and evaluations of top recommended businesses. Figures as of June 2026.

A Note on Method

The factors in this book were derived by measuring and scoring the presence and prominence of each factor across the businesses that AI systems were observed to recommend. They were not obtained from any AI system's stated requirements, nor from any other source claiming to know what an AI needs in order to recommend a business. The relationships described here are observational — they reflect what the data showed, not a proven cause.

Nothing in this book is a representation, warranty, or guarantee that any factor, or any combination of factors, must be present for an AI to recommend a business, or that acting on these factors will produce any particular ranking, visibility, or result. Any suggestion that a given factor is required for an AI recommendation is coincidental to the underlying data and should not be relied on as such.

AI systems change their behavior often and without notice, and outcomes will vary by business, industry, and market. This material is provided for informational and educational purposes only and is not professional, legal, or business advice. Do your own investigation of what ranks a local business in AI applications before making decisions for your business or your clients.

The author is the founder of the platform whose scanning technology and data resources were used to produce the index.

Every active factor in the standard, grouped by category and ordered by weight — what each one measures and how to fix it. This is the complete reference behind the twelve categories. The weights shown are the current monthly read and shift over time.

Business Identity Signals

Local SEO & GBP Signals

Reputation Signals

LSA Eligibility Signals

Contact Prominence

UX & Conversion Signals

SEO

AI Visibility

Content Quality & Expertise Signals

Social Presence Signals

Trust & Safety Signals

AI Corroboration