How to Get Your Business Recommended by ChatGPT, Claude, and Perplexity

By GetGood · July 7, 2026 · 6-minute read

Why this is suddenly worth your attention

More customers now open ChatGPT, Claude, or Perplexity and ask a question that used to go to Google: "Who should I hire to…", "What's the best… near me?", "Is [company] any good?" The assistant reads across the web and answers in a paragraph — often naming two or three businesses and skipping the rest. If you're named, you're in the customer's shortlist before they've clicked anything. If you're not, you may never enter the conversation.

People have started calling this AEO (answer-engine optimization) or GEO (generative-engine optimization). The names are new; most of the work isn't. An assistant is trying to give a person a trustworthy answer, and it favors businesses that are easy to read, clear about what they do, and consistent enough to be believed. That's the same thing that has always made a business easy to recommend — now a machine is doing the recommending.

Can you pay to be recommended?

Not for the organic mention inside a normal answer. When an assistant names a business in the body of its reply, that's the model's judgment, not a slot you purchased — and anyone promising you a guaranteed spot is selling access they don't have. Some assistants are beginning to show separately labeled ads alongside their answers; that's a real advertising product, but it's not the same as being recommended, and customers can tell the difference. Treat "guaranteed AI ranking" the way you'd treat "guaranteed #1 on Google" — as a signal to keep your wallet closed.

What actually makes an assistant name you

No one outside these companies has the exact recipe, and it changes. But the pattern across assistants is consistent, and it's the honest short list worth acting on:

1. Your pages are readable by a machine. The most common reason a business is invisible to AI is dull, not dramatic: the important text only appears after JavaScript runs, and many AI readers see a blank page. If your site reads fine with scripts and images turned off, an assistant can read it too.
2. You answer the question directly, up front. Assistants quote pages that state the answer in the first paragraph, then explain. A page that makes the reader hunt gets skipped for one that doesn't.
3. Your facts are identical everywhere. Same business name, address, phone, and description on your site, your Google Business Profile, and directories. Assistants cross-check; contradictions read as "can't verify this," and unverified businesses get left out.
4. Other credible pages mention you. An assistant rarely trusts your site alone. Genuine third-party signals — reviews, a local news mention, an industry directory, a real customer's post — are what let it say your name with confidence.
5. Your claims are honest. This isn't a moral footnote; it's mechanical. Assistants compare sources, and an overclaim that another page contradicts makes the whole page less quotable. Honesty is the strategy, not the disclaimer.

Notice what's not on that list: no secret AI code, no keyword stuffing, no gaming. The businesses assistants recommend are, boringly, the ones that are genuinely clear and genuinely trustworthy.

How is this different from regular SEO?

It overlaps more than the new acronyms suggest, but three differences are worth naming:

Because the assistant may answer without sending a click, the mention itself is the prize — which is exactly why being described accurately matters as much as being described at all. A wrong recommendation (old hours, a service you dropped, a town you don't cover) can cost you the customer before you ever hear from them.

A one-afternoon starting checklist

You don't need a project. You need a readable, consistent, honest presence. In priority order:

• Confirm your key pages are plain-HTML readable — what you do, where, and for whom, in text a machine can see. This is the highest-leverage fix.
• Put the answer first on each page: a one-line, plain statement of what you offer and who it's for, before the marketing.
• Reconcile your facts across your site, Google Business Profile, and any directory that lists you. One name, one phone, one description.
• Add standard schema.org markup (Organization, LocalBusiness, FAQ) so machines don't have to guess.
• Add an llms.txt file — a plain-text map of your site for assistants. Cheap, honest, and it can only help the tools that look for it.
• Publish a few pages that answer real customer questions in plain language. Every genuinely useful answer is another page an assistant can quote.

How to check whether it's working

There's no dashboard, and that's fine — the honest test is free. Once a month, open ChatGPT, Claude, and Perplexity and type the exact questions your customers would ask: "best [your service] in [your city]," "who should I hire to [the thing you do]," "is [your business] legit?" See whether you're named, and whether the description is right. Write down what you find. The real KPI is offline: customers who mention they found or checked you through an AI assistant. If nothing moves after a couple of months of readable, consistent, honest pages, add more real answers — not more tricks.

Our bias, disclosed: we help businesses set this up, and when we install an AI office we use Claude Code because it can read a business's own files and act on them. None of the checklist above requires us or that tool — a careful owner can do all of it in an afternoon, and we'd rather you know that.