AI Can’t Recommend a Company That Can’t Explain Itself
Search is changing fast. The fix isn’t a new tactic. It’s the oldest one in marketing: knowing exactly who you are and saying it the same way every time.
Try something before you read the rest of this.
Open ChatGPT, Claude, Gemini or Perplexity. Type:
“What does [your company] do, who is it for, and why would someone choose it over [your closest competitor]?”
Read the answer slowly.
Is that you?
Is it the version of you that you’d put on the homepage? Or is it a blurry average of your old website, a directory listing from 2019, and whatever a few reviewers happened to say?
For a growing share of your future customers, that answer is your first impression.
The front door moved
For two decades, the deal with search was simple. You earned a spot on the results page, someone clicked, and they landed on your website, where you got to tell your story.
That deal is breaking.
In a Pew Research Center study of real Google searches, people clicked a traditional search result on 8% of visits when an AI summary appeared. When there was no summary, they clicked on 15% of visits. Almost nobody (1%) clicked the links inside the summary itself.1
These summaries aren’t rare anymore. One large study found AI Overviews on about a quarter of Google searches in the first quarter of 2026.2 Back in 2024, Gartner predicted traditional search volume would fall 25% by 2026 as people moved their questions to AI chatbots.3
So more and more often, the answer gets written before anyone reaches your website.
Which raises the question that matters:
Who’s writing it?
It’s not an AI problem. It’s a clarity problem.
The common reaction from small teams is some version of panic. We need an AI strategy. We need to learn GEO. We need someone who understands prompts.
I get it. But I think that’s the wrong starting point.
AI search doesn’t invent your positioning.
It averages it.
These systems build their answer out of what the web says about you: your site, yes, but also your listings, your reviews, your partners, the local paper, the podcast you did three years ago.
When Ahrefs studied 75,000 brands, the factor most closely tied to showing up in Google’s AI Overviews was branded web mentions (how often and where you’re talked about), well ahead of backlinks.4 Search Engine Land’s 2026 guide put it bluntly: AI engines “strongly favor earned media,” meaning other people’s descriptions of you over your own.5
So here’s the mechanism in one line:
Clear identity → consistent language → other people repeat it → AI repeats it.
When a company is clear, that chain works in its favor.
When it isn’t, everything gets diluted. If your homepage says one thing, your LinkedIn says another, your Google profile describes the business you were five years ago, and your customers describe you three different ways, the model does the only thing it can.
It blends all of that into something generic.
And generic is exactly what doesn’t get recommended.
This was always the job
Nothing about this is new. It’s the same problem marketing has always had, with the consequences sped up.
A while back, at Boundless, my previous consultancy, we worked with a well-known maker of safety products for the outdoors. Their website had grown to 87 disjointed pages. It was trying to talk to distributors and end consumers at the same time. And the copy read like the scientists had written it: accurate, technical, and hard for a normal person to follow.
We rebuilt it around one audience: the person actually using the product outdoors. Distribution got its own site. Eighty-seven pages became about ten. Four clear product categories. Plain-language benefits. Content organized as why first, how second, what it means for you third.
Engagement rose seventeenfold. Over the next twelve months, traffic averaged 175% above the year before.6
That was before AI search existed.
But look at what we did: one clear audience, a small number of clearly named categories, and plain-language answers to the questions real people ask.
That’s almost exactly what AI search rewards today.
The tools changed. What makes a company easy to understand didn’t.
What a lean team should actually do
You don’t need a new department for this. You need a few disciplined habits, and most of them are messaging work, not technical work.
Write the one sentence. Who you serve, what you do for them, and why it matters. If three people on your team would answer that three different ways, AI will too. This is the work under everything else, and it’s the easiest part to skip.
Say it the same way everywhere. Your homepage. Your About page. Your Google Business Profile. LinkedIn. Directory and marketplace listings. The boilerplate at the bottom of your press releases. Consistency used to be a nice-to-have for brand teams. Now it’s how machines figure out who you are.
Answer real questions in plain language. The questions customers actually ask on sales calls and at the counter deserve their own clear answers on your site, under headings that sound like the question. AI systems lean heavily on clear question-and-answer pairs.5 So do people.
Give other people the words. Most of what AI knows about you comes from someone other than you: reviews, partners, local press, associations, podcast hosts. Make your one sentence easy for them to repeat. Ask happy customers for reviews that describe what you actually do, not just “great service.”
Check the plumbing once. Make sure your website isn’t accidentally blocking the crawlers AI tools use (ask whoever manages your site to check the robots.txt file). It’s a five-minute check. It’s also the only purely technical item on this list.
Run the test every quarter. Ask the AI tools the question from the top of this article. Watch whether the answer gets sharper over time.
Fewer clicks isn’t the whole story
One more thing, because it connects to how you measure all this.
Clicks from search are going down. But the people who do arrive from AI tools tend to arrive already informed. Semrush found that AI search visitors are worth about 4.4 times as much as the average traditional organic visitor.2
If you’re still judging your website by raw traffic, this shift is going to look like failure even when it isn’t.
The goal was never just to be found.
It’s to be the answer.
The oldest advantage still wins
A lot of new vocabulary is going to get sold to small businesses over the next few years. GEO. AEO. AI visibility audits. Some of it will be useful.
But underneath all of it is one old idea.
The companies AI recommends won’t be the ones with the cleverest prompts. They’ll be the ones that know who they are, who they serve and why it matters, and that say it the same way everywhere.
That’s always been what earns a recommendation from a person, too.
So before you hire anyone to optimize for the machines, run the test at the top of this page. If you don’t like what comes back, the fix probably isn’t technical.
It’s probably the sentence you haven’t written yet.
Sources
- Pew Research Center, “Google users are less likely to click on links when an AI summary appears in the results,” July 22, 2025 (68,879 searches by 900 U.S. adults, March 2025).
- Contently, “AI Overviews and Organic Traffic: What the 2026 Data Actually Shows,” April 27, 2026, summarizing a 21.9-million-search study (25.11% of Google searches triggered an AI Overview in Q1 2026) and Semrush data (AI search visitors 4.4x as valuable).
- Gartner, “Gartner Predicts Search Engine Volume Will Drop 25% by 2026, Due to AI Chatbots and Other Virtual Agents,” February 19, 2024.
- Louise Linehan and Xibeijia Guan, “An Analysis of AI Overview Brand Visibility Factors (75K Brands Studied),” Ahrefs, May 26, 2025. Branded web mentions correlated at 0.664; backlinks at 0.218.
- Search Engine Land, “Mastering generative engine optimization in 2026: Full guide,” February 23, 2026.
- Boundless Consulting client results, 12 months following the relaunch. See the Proof page.
Decided what your business is for, and can't get it said clearly yet?
Thirty minutes, no pitch. Tell me what you're working on and I'll tell you honestly whether the problem is marketing at all — and whether I'm the right person for it.