How to test your brand's visibility on AI engines

Testing brand visibility in AI engines requires a protocol. Asking ChatGPT once if they know the company only measures one response at that moment. A useful test uses queries close to the purchasing journey, repetitions, and defined criteria before looking at the result.

Build the query universe

Separate the questions into four groups: problem, category, comparison and decision. Include persona, segment, location, and restriction variations. For a B2B company, “financial software” and “financial software for industry with ERP integration” represent different disputes.

Avoid brand name prompts in the discovery phase. Then, test recognition and accuracy: what the company does, for whom, what products it offers and how it differentiates itself.

Standardize the test

Record available platform, model or mode, date, language, location, login state, and web search usage. Run each question at least a few times and preserve the complete answer with sources. Variation is part of the phenomenon; don't just select favorable execution.

Measure distinct results

Track mention rate, citation rate, share of response among competitors, position of first mention, description stickiness and sentiment. Also mark non-searchable answers and third-party sources. A brand may be remembered for its model, but not supported by a current source.

Use platform data

In 2026, Search Console began providing a generative AI performance report for a subset of properties, with AI Overviews and AI Mode impressions by page, country, device, and date. Bing Webmaster Tools also now shows citations in AI responses, including URLs and grounding queries.

These reports complement manual testing. They don't cover every platform or explain every answer. Combine them with benchmark analytics, assisted conversions, and brand monitoring.

Compare versions over time

Create a baseline, make identifiable changes, and repeat the protocol. Do not attribute improvement to a change if indexing, news, model or competition have also changed. The goal is to find sufficiently stable standards to guide content and technology.

AI visibility diagnostics helps you start this analysis with a website and a priority query.