What changes when buyers ask ChatGPT instead of Google
AI answers name one or two companies and send no click. Your analytics cannot see any of it.
The short answer
When a buyer researches inside an AI engine rather than a search engine, they get a synthesised answer naming one or two companies rather than a page of links — and no click is sent, so the interaction never appears in your analytics. Being cited depends on whether an engine can parse your content, verify your claims and identify you as an entity, which is a structural property of your site rather than a ranking you can bid for. The practical consequence is that a growing share of consideration now happens somewhere you cannot currently measure.
The mechanics are different, not just the interface
A search engine returns ranked documents and the user chooses. An AI engine returns a composed answer and the user usually stops there. Ten blue links becomes one paragraph naming two companies.
That compresses the funnel to almost nothing. There is no page two, and being the third-best answer is often indistinguishable from being absent.
You cannot see it happening
No click means no referrer, no session, no attribution. A buyer can research your category, be given a competitor as the recommendation, and form a shortlist without generating a single event in any system you own.
This is why the shift is consistently underestimated. It does not show up as a decline. It shows up as pipeline that never existed, which nobody investigates.
What actually determines whether you get named
- Whether an engine can parse your content at all — structure, headings, and whether the substance is in text rather than locked in images or rendered late.
- Whether your claims are verifiable against other sources it trusts.
- Entity clarity — whether it can tell reliably who you are, what you do and where, without confusing you with a similarly named company.
- Whether your content is shaped like an answer to a question somebody actually asks.
None of this is a ranking you can buy. It is closer to technical SEO than to advertising, and it responds to the same kind of structural work.
The reputational half nobody checks
Engines synthesise from whatever is public and parseable. Where your own record is thin, they fill the gap — with a superseded price, a discontinued product, a comparison you never made, or a claim from a forum.
That is a live commercial exposure with no owner, no monitoring and no remediation path, repeated to every buyer who asks. Most organisations have never checked what is being said about them, because there is no inbox it arrives in.
What to actually do about it
- Measure first. Establish how several engines describe you today against your named competitors, and where they cite someone else.
- Fix the structural causes — entity clarity, parseable substance, answer-shaped content, verifiable claims.
- Correct what is wrong. False statements are fixable, but only once you know they exist.
- Then keep watching, because the answers change as the underlying models and their sources do.
Why we can measure this cheaply
We built and operate GenAI Ranker, a multi-tenant platform that tracks how six AI engines describe a brand against its competitors. Running a scan costs us very little, which is why we run the first one free rather than selling it.
It is also the most direct way to find out whether any of this applies to you. For plenty of categories the honest answer is that buyers are not researching there yet — and that is worth knowing before anyone spends money on it.
Questions this raises
Is AI search visibility just SEO with a new name?
They overlap and they are not the same. Classic SEO optimises for ranking in a list of documents; AI visibility is about being the source an engine composes its answer from, which depends more on entity clarity, verifiability and answer-shaped structure. The technical foundations help both, but you can rank well in Google and still be invisible inside an AI answer.
Can we pay to appear in AI answers?
Not in the way you buy search ads. Citation follows from being parseable, verifiable and clearly identifiable, which is structural work on your own content and data rather than a bid. That is the good news — the advantage compounds and does not stop when you stop paying.
How would we know whether this matters for our category?
Measure it. If engines are already naming competitors when someone asks about your category, it matters now. If the category is not being researched that way yet, it does not, and you should spend the budget elsewhere — which is an answer worth having either way.
This is what we do about it
More like this
What actually breaks an AI agent in production
The failure modes, in the order we have met them — and almost none of them are the model being wrong.
How to cost an AI automation before you build it
Four lines of arithmetic that decide whether a project is worth funding — and that most proposals skip.
Why your RAG system should refuse more often
A retrieval system that never says “I don’t know” is not accurate. It is unmeasured.