How to get recommended by ChatGPT

A buyer opens ChatGPT and asks which provider is best for their situation. One or two names come back, and a deal starts forming around a company that never knew the conversation happened. Whether yours is the name returned is not luck, and it is not something you can buy. It follows a mechanism most teams are quietly working against.

That is the short answer. The mechanism is specific, and most teams have never looked at it.

How do I get ChatGPT to recommend my company?

Get recommended by ChatGPT by becoming the name the model most associates with your category. It names the providers that appear most consistently in trusted, well-structured content about the problem. Publish clear answers to your buyers' questions, then earn the credible third-party references and structured signals that let the model read and attribute you.

Why ChatGPT names who it names

ChatGPT does not hold an opinion about your market. It predicts the most probable answer to a question, based on patterns in the text it has read. When someone asks for the best provider in a category, the model returns the names that co-occur most often with that category in credible sources, expressed with enough confidence that the pattern is strong.

Two inputs shape this. The first is the training data, the large body of text the model learned from. The second, in tools that browse, is live retrieval, where the model pulls current pages to ground its answer. Both reward the same thing. A company referenced widely and consistently, in language a machine can parse, becomes the probable answer. A company mentioned rarely, or only on its own site, does not.

This is why the recommendation hardens over time. Once a name appears in the answer, more people cite it and link to it, which reinforces the pattern the model learned. The first credible mover in a category accumulates references while everyone else is still deciding whether the channel is real. The answer settles around them, and resettling it later is slow work.

Is this just SEO with a new name?

No. The two share foundations and divide on the job they do. SEO earns a position in a list of links, measured by ranking and clicks. Getting recommended by ChatGPT earns your name inside a generated answer, measured by whether the model names you. You can hold the top organic ranking and still be absent from the answer a buyer reads above your listing, because that answer is assembled, not ranked.

The overlap is genuine, so concede it plainly. Crawlable, clearly structured pages with authoritative content help both. A page the model cannot read or attribute will not be cited, the same way a page Google cannot crawl will not rank. What you optimise for is where they part. Search rewards the page that best matches a query. A model rewards the name that most reliably belongs in the answer, a function of how often and how credibly that name appears across everything the model has read, rather than the merit of any single page.

Seer Interactive found in 2025 that organic click-through to the top result falls by around 60% once an AI answer sits above it. Pew Research Center reported in 2025 that when Google shows an AI summary, people click through to a site just 8% of the time. The ranking can be intact while the traffic it used to send is gone. That gap is why the recommendation is now the thing to win, even where the ranking still holds.

The sequence that gets you named

Getting recommended is a build, run as a sequence rather than a single tactic. Each step feeds the next, and the order matters.

Step	What you do	Why it works
Audit	Ask AI the questions your buyers ask and record where you stand, question by question	You cannot improve a position you have not measured; most teams have never seen theirs
Content	Publish clear answers to those buyer questions in the format models pull from	Gives the model trustworthy, parseable material that names you in context
Citation building	Earn third-party references and structured signals that point to your name	Recommendations rest on what others say about you, not only what you say about yourself
Re-scoring	Re-run the question set monthly and track the movement	Shows which questions you now own and where a rival is still the answer

Start with the audit because it sets the target. Run your real buyer questions through ChatGPT and the other engines, and record where the answer names you and where it names a rival. The questions that sit open, with no clear leader, are usually where the fastest wins are, because no model has settled on anyone yet.

Then build the content the model can draw on. Write direct answers to the questions buyers actually ask, in plain structured prose, with the claim near the top and the supporting detail beneath. A model assembling a recommendation pulls from material that states things clearly and can be lifted out cleanly. Marketing pages built around slogans give it nothing to quote.

Citation building is the part SEO habits underprepare you for. The model trusts your name partly on the strength of who else uses it. References from credible third-party sources and consistent, structured descriptions of what you do across the web teach the model to attribute the category to you. This is the slow compounding layer, and it is where early movers pull ahead.

Re-scoring closes the loop. Run the same questions every month and score the change. The metric that matters is your named-in-answer rate across the questions your buyers ask, tracked over time. It tells you whether the work is landing and where to point it next.

What you cannot do

You cannot pay ChatGPT to recommend you. There is no advertising slot inside the answer and no sponsored list you buy onto. Anyone selling a shortcut into the recommendation is selling something the mechanism does not contain.

You also cannot fake the references. Thin pages stuffed with your name, or claims the rest of the web does not support, tend to weaken trust rather than build it, because the model weighs consistency across sources. The recommendation rewards a real, verifiable presence. That is slower than a trick, and it is also far harder for a competitor to dislodge once you hold it.

Being early is the lever you do control. The category for AI recommendations sits open in most B2B markets because few teams are working on it yet, and every month a rival is the answer, the model settles harder on them. The cost of starting is lowest now.

Does this actually win deals?

It already has. One healthcare compliance platform we work with saw its largest contract of the period arrive through ChatGPT, after a buyer asked the engine for a recommendation and acted on the name it returned. The AI audience for that platform grew 36% month on month, and qualified pipeline moved from £240k to £645k over the programme. Names are withheld by agreement; the outcomes are confirmed.

The point is not the single contract. It is that the buyer never searched and never compared a page of links before the shortlist formed. The introduction was made inside the chat, by a model that named one company and left out the rest. Being that company is the work.

How Forge does this

At Forge Together we run this as a programme called generative engine optimisation, in the four steps above: audit, content, citation building, and monthly re-scoring. We start by showing you where you stand, then build your name into the answers your buyers are already asking. For the wider method, see the B2B GEO playbook and what generative engine optimisation is. If you want the content half in more detail, read how to get cited in Google AI Overviews and AI answers.

The teams scoring their position now are the ones who will hold the answer when the rest of the market arrives.

Book a discovery call and we will run a free AI-visibility audit at forgetogether.agency/GEO. 30 minutes. No pitch deck. We'll run your category live on the call and show you exactly what AI says about you today.

FAQ

How do I get ChatGPT to recommend my company?

ChatGPT recommends the provider that appears most consistently across trusted, well-structured content about a category. Answer your buyers' real questions in content the model can parse, then back it with third-party references that confirm who you are. Track your named-in-answer rate monthly to see whether the model is starting to name you.

Can I pay to be recommended by ChatGPT?

No. There is no advertising slot or paid placement inside a ChatGPT recommendation. The model names providers based on how often and how credibly they appear across the content it has read and retrieves. Anyone offering to buy you into the answer is selling something the mechanism does not contain.

Is getting recommended by ChatGPT just SEO?

No, though they share foundations. SEO earns a ranked position in a list of links. Getting recommended earns your name inside a generated answer. You can hold the top organic ranking and still be absent from the AI answer a buyer reads first. Winning a citation takes different work from winning a click.

How long does it take to get recommended by ChatGPT?

It varies by category. Questions that sit open, with no provider named yet, can move within weeks once you publish strong content and structured signals. Categories where a competitor is already the default answer take longer, because the model has settled on a pattern that consistent references must shift over months.

How do I know if ChatGPT recommends my company?

Run the questions your buyers ask through ChatGPT and the other engines, and record where the answer names you against where it names a rival. Repeat monthly and track the change. Your named-in-answer rate across that question set is the measure of whether AI recommends you, scored over time rather than once.