Frequently asked questions

What is GEO (Generative Engine Optimization)?

GEO is the practice of improving how AI assistants (ChatGPT, Gemini, Claude, and similar systems) describe and recommend a company when people ask them questions. Where SEO targets search rankings, GEO targets the answers themselves: whether they're accurate, whether they cite your sources, and whether they mention you at all. It matters because AI assistants are increasingly the first place buyers form an opinion about a product.

How is GEO different from SEO?

SEO gets your pages ranked in a list of links; the visitor still reads your site and forms their own view. With AI assistants there is often no click: the assistant synthesizes an answer from many sources and the buyer takes that answer at face value. GEO therefore focuses on the substance and sourcing of the answer: correcting inaccuracies, making your own material the material the engines draw on, and being present when your category comes up. Good SEO helps GEO, but ranking well and being described accurately are different problems.

How do you check what AI assistants say about a company?

We ask the major assistants the real questions a company's buyers ask about the product, its capabilities, and the category, and record the full answers. Then we check each answer against the company's own source of truth: documentation, published claims, actual certifications. The output is a set of answers annotated with what's wrong, what's missing, and which sources the assistants relied on instead of the company's own.

What does the free Snapshot include?

You send us your company name, a short description in your own words, and about five real buyer questions. We put those questions to ChatGPT, Gemini, and Claude with web search on and place the answers side by side. An automated pre-pass highlights suspect claims and flags competitor sources in red; then you check the answers against what you know to be true. No call required, no obligation.

Is the Snapshot really free?

Yes. It's free for the first 100 companies, as a launch offer; after that it's 9€, then 19€. Either way, its cost is credited in full toward any later audit, and refunded if no material issue is found. The idea is simple: you should only pay if there's genuinely something to fix.

How much does it cost?

The offer steps up in tiers: the Snapshot (free at launch), the Focused Audit at 390€, the Expert Accuracy Audit at 1,490€, the Guided Fix Sprint at 3,900€, and Ongoing Monitoring at 290€ per month. Each tier only makes sense if the previous one showed something worth acting on. Full details are on the pricing page.

Who is this service for?

Founders, product leads, and communications leads at technical and deep-tech companies: businesses with complex or specialized products where getting the details right matters. It's most valuable when your product's claims are precise (certifications, capabilities, limitations) and a confident-but-wrong AI answer could cost you a deal. You don't need any marketing background; the work is designed for people who know their product, not people who know advertising.

Can you guarantee AI assistants will recommend my company?

No, and you should be wary of anyone who says yes. Nobody controls what these models say; the engines are updated constantly and their answers are synthesized from sources no single party owns. We sell the measurement and the process, not a sentence that ChatGPT will say on command. What can be done, and measured, is improving the inputs: making your own material clear, machine-readable, and authoritative enough that the engines draw on it, then re-testing to confirm the answers improved.

What happens to our data, and are your findings legal advice?

To test a product, we send the company name and public information to OpenAI, Google, and Anthropic, which run ChatGPT, Gemini, and Claude; that's inherent in querying these engines. We don't need your confidential data to do it. Our findings compare engine output against your public source of truth: they're visibility observations, not regulatory or legal advice. At the Snapshot tier, you adjudicate what's true; we show you the raw answers and the automated flags.

Why does the person auditing need domain expertise?

Because the errors that matter are subtle. An assistant describing a screening tool as a diagnostic one, or a research-use capability as certified, produces an answer that looks perfectly fine to a generalist. Judging whether an AI's answer about a technical product is actually correct requires understanding the product's field, which is why our audits evaluate substance, not just whether your brand appears.

What happens after an audit finds problems?

The audit ranks the gaps by how much they'd affect a buying decision. From there, we guide your team through the fixes (changes to your site, documentation, and the third-party material the engines rely on) rather than doing it for you, because your team understands the product best. We measure the assistants' answers before and after, so improvement is demonstrated, not assumed.