AI visibility is an official category now. Most tools study AI's answers and guess backward. We check what AI reads about you before it answers.
BLUF: AI visibility is an official software category now. Gartner published a buyer's guide for it in March, and on September 15 one of the companies in it raised $180 million. We're in that category, and I'm fine saying so. We work from the other end of it, though. Most tools study the answers AI gives and try to guess backward to what caused them. We check what AI reads about your business before it answers, using a list of things buyers have needed to see for 20 years. The answers keep changing. That list hasn't.
Yes. For the last two years, "AI visibility" was mostly a phrase in cold emails. That's over.
In March, Gartner (the research firm big companies pay to tell them what software to buy) published its Market Guide for Answer Engine Visibility Tools. Then on September 15, Profound raised $180 million at a $1.8 billion valuation, with Sequoia Capital and Kleiner Perkins leading the round. Profound says it serves more than 1,000 enterprise brands.
When Gartner writes a buyer's guide and Sequoia writes a check that size, the argument about whether this category exists is finished.
Marcus Sheridan and I went back and forth on this earlier this month. Where we landed: we're in this category. Buyers are going to file us there anyway, so we might as well say it out loud. What matters is how we work inside it, and that comes down to which end of the black box you look at.
Ask ChatGPT who the best roofer in town is, and a lot happens between your question and its answer. The model searches, reads, weighs sources and writes. Exactly how it weighs those things is known only to the people who build it. And they change it all the time, usually without telling anyone.
So it's a black box. A question goes in, an answer comes out, and the middle is private.
It's a moody black box, too. Mike Sonders ran 12 buyer questions through ChatGPT 100 times each and wrote it up for Search Engine Land. Across 100 runs of a single question, ChatGPT named about 44 different brands. Any one answer named about 10. Only about 5 brands per question showed up 80% of the time or more.
Mike King at iPullRank put it well in a piece this month: "For most brands in most categories, whether you 'appear' for a given prompt on a given day is closer to a dice roll than a ranking."
Most of them work the same way. They run lots of buyer-style questions through ChatGPT, Gemini and the rest, collect every answer, and count how often your name comes up. When we read the homepages of 32 companies selling these tools, effectively all of them led with that number.
Then they work backward. You showed up less this month, so maybe it's your reviews. Or your website. Or a competitor's new article.
Think of a restaurant. These tools are food critics. They order the dish, taste it, score it, and guess what the chef did in the kitchen. The trouble is the chef changes the recipe every few weeks and never publishes it. And as Sonders showed, the same dish ordered twice can come out different.
That doesn't make tasting useless. King is careful on this point: you can measure AI answers well if you ask many times with the same method and report a range instead of one tidy number. Even a well-measured score tells you what came out, though. It can only guess at why.
And there's a blind spot built into that approach. Much of what AI read before it answered, the pages and reviews and listings it weighed, never shows up in the answer. A few sources get a link underneath; the rest stay hidden. You can't see all of it by reading answers, no matter how many you collect.
That's the side we work on.
AI Trust Signals starts on the other side of the box. Before AI can recommend your business, it has to read about your business: your website, your reviews, your pricing page, what other sites say about you. Those are the ingredients. We check whether they're there, how good they are and what's missing.
How do we know which ingredients matter? We didn't reverse engineer them from AI answers. Marcus has taught them for 20 years. His book, They Ask, You Answer, boils it down to 5 things buyers want to know before they trust a business: what it costs, what can go wrong, how you compare, what customers say, and who's best. Businesses that answered those questions openly earned buyer trust long before AI showed up. (We walked through all five in AI Keeps Moving the Goalposts.)
Here's the key idea. AI wants the same things your buyer wants, for the same reason. An AI company lives or dies on whether people trust its recommendations. If it sends someone to a bad roofer, that person stops asking. So AI goes looking for the same proof a careful buyer looks for: clear prices, real reviews, named people, plain answers about problems.
The gatekeeper changed. The buyer didn't. AI caught up to Marcus.
This matters most when you're deciding where to spend your time and money.
The inside of the box will keep changing. New models, new search features, new rules about which sources count. Every change shuffles the answers, and every tool that works backward from the answers has to start guessing again.
What AI reads for moves far less. Clear prices were a reason to trust a business in 2006. They're a reason in 2026. They'll still be one whenever the next model ships. Work you put into your ingredients keeps paying off no matter what the chef does to the recipe.
How AI reads your site does change, though. Which AI tools can get into your site, how fast your pages load, how your facts are labeled for machines. That's why part of the AI Trust Signals Score asks a simple question: can AI read you? Those basics decide whether AI can read your ingredients at all. What it looks for once it can read them has stayed the same.
The free check shows you what AI says when a buyer asks about your kind of business. That's the mirror, and you should see it. Better you hear it from AI than from a salesperson.
Your fix comes from the ingredients, though. The AI Trust Signals Score rolls 19 signals AI checks into one number, and each signal comes with the fix.
The ingredients tell a clear story on their own. The average business we score lands at 61 out of 100. On our scale, that's Invisible: AI can't see the business well enough to vouch for it yet. Every signal behind that number is something you can work on.
If you're weighing AI tools, ask each one a single question: when AI changes how it works next month, what will this tool tell me to do differently? If the answer is "we'll know when the answers change," you're paying someone to chase the chef.
The work that lasts is the work buyers have always rewarded. Put your prices where people can see them. Talk openly about what can go wrong. Collect reviews on the sites AI reads. Show the real people behind your business.
People still buy from businesses they trust. Now they ask AI first.
See what AI says about your business right now. Run the free check at aitrustsignals.com/recommends. No signup, and you'll have answers in minutes.
AI visibility is how often, and how well, AI tools like ChatGPT and Google's AI mention your business when people ask questions. Most of the software that tracks it works by asking AI many questions and counting how often your business appears in the answers.
It can tell you where you stand, and it's worth seeing once. Keep in mind that AI answers change from one ask to the next, so any single score is a snapshot. Tracking tells you what AI said. It can't tell you for sure why, and the "why" is where your fix lives.
The same things a careful buyer checks: clear prices, honest answers about problems, how you compare to others, real customer reviews, and proof of real people behind the business. AI companies need people to trust their recommendations, so they look for the same proof people do.
Most AI visibility tools study the answers AI gives and work backward. AI Trust Signals checks what AI reads about your business before it answers, scores it across 19 signals AI checks, and hands you the fix for each one. We look at the answer too, through our free check, but the fix comes from the ingredients.
That's the reason we built it this way. AI companies change how their tools work all the time, which shuffles the answers. What buyers need to see before trusting a business has stayed steady for 20 years, and AI looks for the same things.
See what AI says about your business.
