How Founders Get Their Brand Recommended by AI: What Changed for ChatGPT, Perplexity, and Google's AI Mode in 2026
For founders and marketing leaders who want their company to show up when buyers ask an AI tool who to hire or what to buy.
Try this before you read any further. Open ChatGPT, Perplexity, or Google's AI Mode and ask the question your best customer asked before they found you, something like "who are the best firms for this kind of problem" or "what should I look for in this kind of product." Read the answer slowly. Look at which companies are named, how they are described, and which websites the tool cites as its sources. For a lot of founders, that answer is the first time they realize that a buyer can form a complete opinion about their category without ever seeing their website.
When the first version of this post went up, we said that AI tools build their answers from training data rather than from keywords and backlinks, and that they were weak at local and real-time information. That was a fair read of the tools at the time, and it is only half right now. The major assistants search the live web, cite their sources, and pull from reviews, forums, news, and business listings when they answer. So the work has shifted. Training data still shapes what an AI tool already "knows" about your brand, but what it finds and cites when someone asks today matters at least as much.
The good news for newer companies is the same as it was then. You do not need to be a household name to be recommended. You need the right people, in the right places, saying clear and consistent things about you.
Your own website is a small part of what AI tools read about you
Most founders start with their site, and the data says most of the answer comes from somewhere else.
The most common mistake I see is treating AI visibility as a website project. Founders rewrite their homepage, add a few keywords, and wait for ChatGPT to notice. The site matters, but it is a smaller input than most people assume.
McKinsey's State of the Consumer 2026 report, published in June, looked at which sources large language models cite when answering consumer goods questions. Based on citation data from October 2025 to May 2026, only 1% of cited sources were brand-owned websites. Even among the ten most-cited sites, which together account for about 22% of all citations, brand websites made up no more than 10%. The rest came from forums, retailer reviews, blogs, video platforms, and publishers. That study covers consumer goods, so the exact share will differ in B2B and professional services, but the direction is hard to ignore.
McKinsey also named a problem I have watched play out with real brands. When the sources an AI tool reads describe a company in inconsistent or incomplete ways, the report calls it "signal dissonance," and AI models are less likely to include that brand in an answer or to describe it accurately.
An everyday comparison helps here. Asking an AI tool for a recommendation is a lot like asking a well-read friend where to eat. Your friend repeats what they have heard from many people. If half the people they trust call a restaurant a great taco spot and the other half call it a sushi bar, your friend will probably recommend somewhere else, because they are not sure what that place actually is.
The three signals AI tools look for: consistent, corroborated, and current
A simple way to organize the work, with the reason behind each piece.
I group the work into three signals, because almost every useful tactic falls into one of them, and it keeps teams from chasing every new trick that shows up on social media.
1. Consistent: say the same thing about yourself everywhere. Write one clear description of what your company does, who it serves, and what makes it different, then use that same language on your website, your LinkedIn company page, your Google Business Profile, your founder bios, your directory listings, and your press boilerplate. The reason is signal dissonance. Every place your description drifts is a place where an AI tool may form a fuzzier picture of you. This replaces the old advice about repeating keywords. Keywords still help people find you, but the goal now is a stable, recognizable description, not keyword density.
2. Corroborated: get other people to say it about you. This is the piece that matters most, because it is where most of the citations come from. Earned media, guest articles on respected industry sites, podcast interviews, inclusion in credible "best of" roundups, customer reviews, and genuine participation in the forums where your buyers ask questions all count. McKinsey's report recommends the same mix: stronger public relations, consistent messaging across platforms, and an active presence in the forums where consumers already discuss a brand. Gartner's September 2026 consumer survey adds a reason it works. Of 1,006 U.S. consumers surveyed, 43% said they now rely more on real people for shopping information and recommendations because of how much AI-generated content they see. AI tools and buyers are both looking for the same thing, which is someone other than you vouching for you.
3. Current: make your information easy to find, read, and trust today. Because AI assistants now search the live web, freshness and structure matter. Keep key pages up to date, use clear headings, and add plain FAQ sections that answer the questions buyers actually ask. McKinsey points to structured, scannable content such as FAQs and lists, and to publishing detailed material that was not public before, like technical documentation or original research. If you have data or expertise nobody else has, publishing it gives AI tools something worth citing.
AI tools recommend the brands other people describe clearly and consistently, which means your reputation now reaches buyers through pages you do not own.
Getting recommended by AI
The three signals AI tools look for.
AI assistants recommend the brands that are described clearly, by many sources, with information they can find today.
-
Consistent
Say the same thing about yourself everywhere.
One clear description on your site, LinkedIn, Google Business Profile, bios and listings. Mixed signals make AI tools less likely to include you.
-
Corroborated
Get other people to say it about you.
Earned media, reviews, interviews, credible roundups and real forum participation. 43% of US consumers now rely more on real people for shopping recommendations. Gartner, September 2026
-
Current
Make your information easy to find today.
Many AI assistants search the live web. Keep key pages updated, add plain FAQs and publish original data worth citing.
Start today
Ask three AI tools the ten questions your buyers ask most. Record who gets named and which sources get cited.
Local search still matters, and it now feeds AI answers too
The old advice said to balance AI and local. The two now overlap.
The first version of this post said AI tools were weak at local search and that founders should keep local SEO as a separate track. Local SEO is still essential if your customers are nearby, but it is no longer separate. When someone asks an AI assistant for a nearby service, the answer often draws on business listings, maps data, and reviews. That means the basics, a complete and accurate Google Business Profile, consistent name, address, and phone number across directories, and a steady flow of genuine reviews, now do double duty. They help you in traditional local results, and they give AI tools reliable information to repeat. How much each assistant relies on each source varies and changes often, so treat this as a direction rather than a rulebook.
What to stop doing
A few tactics from the earlier playbook no longer earn their time.
Hashtag strategy and posting volume for its own sake are low-value now. Social activity still helps when it starts real conversations or earns mentions, but a steady stream of thin posts does not make an AI tool more likely to recommend you. The same goes for "keyword-rich" blogging. Gartner's September 2026 survey found that 65% of consumers think brands are producing too much AI-generated content, and 57% said that content made them trust brand messaging less. Fewer, more substantial pieces that someone else would want to quote are a better use of the same hours.
What the research shows
AI answers and brands
Brand websites supply 1% of the sources AI cites in consumer goods. Gen Z already shops with it.
Large language models build their answers from reviews, communities and other sites, not from brands. Younger shoppers are turning to those answers first, while consumers say brands produce too much AI content.
| Finding | Figure | Source |
|---|---|---|
| Where AI answers come from | ||
| Share of LLM-cited sources that are brand-owned websites, consumer goods, October 2025 to May 2026 | 1% | McKinsey, June 22, 2026 |
| Brand websites’ share among the ten most-cited sources | up to 10% | McKinsey, June 22, 2026 |
| Change in overall open web traffic since 2023 | down 8% | McKinsey, June 22, 2026 |
| Shoppers · McKinsey survey, five countries | ||
| Gen Z versus baby boomers already using gen AI tools for shopping | 28% vs 16% | McKinsey, June 22, 2026 |
| Gen Z versus baby boomers who regularly use the AI overview at the top of search results | 60% vs 29% | McKinsey, June 22, 2026 |
| Trust · Gartner survey, US | ||
| Consumers who say brands produce too much AI-generated marketing content | 65% | Gartner, September 22, 2026 |
| Consumers who now rely more on real people for shopping information and recommendations | 43% | Gartner, September 22, 2026 |
McKinsey surveyed 4,863 consumers in Brazil, France, Germany, the UK and the US between March 31 and April 8, 2026; its citation figures cover consumer-goods queries from October 2025 to May 2026. Gartner surveyed 1,006 US consumers between May and June 2026. These are survey and citation-sample results: they describe patterns and stated behavior, not guaranteed outcomes for any one brand.
These are survey and citation-sample results. They describe patterns and stated behavior, not guaranteed outcomes for any one brand.
What this likely means over the next 6 to 18 months
Expect AI tools to move closer to the purchase itself.
McKinsey expects gen AI tools to play a growing role in the last stage of buying, where AI agents compare options and complete purchases on a customer's behalf, and it notes that Google is already piloting ways for brands to surface offers directly inside AI recommendations. If that holds, the brands that AI tools already describe clearly will most likely have an advantage when those agents start shortlisting. Nobody outside these companies knows exactly how each system weighs its sources, and the weighting will keep changing, so I would not bet on any single tactic. I would bet on the three signals, because consistent, corroborated, and current information is what any reasonable system should reward.
If you want an outside read on how AI tools currently describe your brand, Camino5 can run that review with your team.
How to start today
Ask ChatGPT, Perplexity, and Google's AI Mode the same ten questions your buyers ask before they hire or buy, and record who is named and which sources are cited. Ten is a practical starting set, not a magic number, and repeating the same questions monthly lets you see change over time.
Write one 50-word description of your company and update your website, LinkedIn, Google Business Profile, founder bios, and directory listings to match it. This is the fastest way to reduce signal dissonance.
Take the source list from step one and mark which publications, forums, and review sites appear most often. Those are your highest-value targets for earned media and genuine participation.
Add a plain FAQ section to your most important service or product page, answering the questions from step one in clear, direct language. This gives AI tools a current, well-structured source to cite.
Ask your five most satisfied recent customers for a review on the platform your buyers already use. Reviews are one of the few corroborating signals you can start building this week.
The founders who get recommended over the next year will most likely not be the ones who found a trick, but the ones whose customers, partners, and press all describe them the same way.