How to get your business recommended by ChatGPT, Claude and Gemini
Search is no longer ten blue links. Your next customer might ask an AI assistant who to hire and never see a results page. Here is how to become the answer.
A client called me last year, half amused and half worried. A new customer had told him, "I asked ChatGPT who the best supplier in our area was, and your name came up." He had never optimised for that. He did not know you could. He just wanted to know how to make it happen on purpose.
That conversation is happening more and more, and most businesses are completely absent from it. They have spent years fighting for Google rankings while a parallel discovery channel quietly opened up beside them, one where the answer is not a list of ten links but a single recommendation. If you are not the recommendation, you do not exist.
To get recommended by ChatGPT, Claude and Gemini, you need to be mentioned consistently across the sources they trust: your own clearly-structured website, third-party reviews, directories and editorial mentions. AI assistants do not rank pages, they synthesise what the web already says about you. Become the obvious answer in that wider conversation and the assistants repeat it.
Why this is not just "SEO for robots"
The instinct is to treat AI search like another ranking game. It is not, and that misunderstanding is why most attempts fail. Google shows you a page. An AI assistant gives an answer, then sometimes cites where it came from. The unit of visibility changed from "a ranking position" to "being part of the answer". That sounds subtle. In practice it changes almost everything about what you optimise for.
When someone asks Claude or ChatGPT "who should I hire for X near me", the model is not crawling the live web for most queries. It is drawing on what it has already absorbed about the entities in that space, supplemented by live retrieval on the platforms that do search the web. It is reconstructing a consensus. Your job is to make sure the consensus includes you, described accurately, in the words you would choose.
The framework I use: be mentioned, be clear, be corroborated
Every time I take on this work, it comes down to three things working together. Miss one and the other two underperform.
1. Be mentioned (entities, not just pages)
AI assistants think in entities: a business, a person, a place, a service, and the relationships between them. Before you worry about content, make sure the basic facts about your business are stated identically everywhere they appear. Your name, what you do, where you operate, who you serve. Inconsistency here is poison. If three sources describe you three different ways, the model has no stable entity to recommend.
2. Be clear (structure the model can lift)
Models reward content they can extract cleanly. That means direct answers to real questions, near the top, in plain language. It means structured data that spells out who you are and what you offer. It means a page that answers "what does this business do and who is it for" without making anyone, human or machine, dig for it.
3. Be corroborated (the web has to agree)
This is the part businesses hate, because you cannot fully control it and it takes time. Reviews, directory listings, mentions in articles, citations from other credible sites. AI platforms lean heavily on third-party corroboration because it is harder to fake than your own marketing. One glowing page on your own site means little. The same story echoed across Google reviews, a directory, and an industry mention is what moves the needle.
The businesses that show up in AI recommendations are almost never the ones with the cleverest content. They are the ones with a coherent, corroborated presence: consistent facts, real reviews, and a few credible mentions. At DigitalRyze the AI-visibility work that produces results is unglamorous. It is mostly making the web tell one clear story about a business, over and over, until the machines repeat it back.
A checklist you can run this week
- Search your own business in ChatGPT, Claude, Gemini and Perplexity. Ask "who is [your business] and what do they do". Note what each gets wrong. That gap list is your roadmap.
- Make your core facts identical across your website, Google Business Profile, and your top three directory listings. Same name, same description, same service list.
- Add or tighten structured data on your key pages so the entity is machine-readable, not just human-readable.
- Write one genuinely useful answer page for the question your best customers actually ask before buying. Answer it in the first paragraph, then go deep.
- Get three more real reviews this month, and reply to them. Recency and volume both matter.
Common mistakes
Treating it as a one-time task. AI training and retrieval refresh over time. A business that goes quiet fades from the consensus. This is maintenance, not a project.
Faking corroboration. Bought reviews and spammy citations are exactly what these systems are getting better at discounting. You are building an asset that compounds, or a liability that eventually gets caught.
Ignoring the basics to chase the shiny part. I have watched businesses obsess over "prompt optimisation" while their Google Business Profile listed the wrong opening hours. Fix the boring foundation first.
The opportunity right now is timing. Most of your competitors have not noticed this channel yet, which means the consensus in your space is still being written. The businesses that show up consistently over the next year are likely to keep that position, because being the established answer is self-reinforcing. The ones who wait will be trying to dislodge an incumbent later. I would rather you were the incumbent.
Frequently asked questions.
How is getting recommended by ChatGPT different from ranking on Google?
Google returns a list of pages and you compete for position. An AI assistant returns a synthesised answer and may recommend specific businesses. You are not competing for a slot, you are trying to become part of the consensus the model has formed about your space, which depends on consistent mentions, clear structure and third-party corroboration across the web.
Can I pay to appear in AI recommendations?
No, not directly, and you should be wary of anyone who says otherwise. AI assistants build recommendations from organic signals: your structured site content, reviews, directories and credible mentions. The work is earning a clear, corroborated presence, not buying placement.
How long does it take to start showing up in AI search?
It varies, but expect it to build over months rather than days. Entity consistency and structured data can be fixed quickly, while the corroboration layer (reviews, mentions, citations) compounds over time. Local and niche queries tend to move faster than broad competitive ones.
Which AI platforms matter most for my business?
Start with the one your customers actually use, usually ChatGPT given its scale, then Gemini because of its tie to Google surfaces, then Claude and Perplexity. Visibility on one does not guarantee visibility on the others, so check each separately rather than assuming they share a view of you.
