AI visibility audit: how to benchmark your company against competitors in ChatGPT, Gemini, Perplexity and Copilot

Budgets for AI visibility often start with a tool subscription or a content push. Without a baseline, nobody can tell later whether the money changed anything, or whether it went to questions where you were already winning.

An audit answers three questions first: where are we named today, who is named instead of us, and which sources are assistants using to decide. Since OpenAI launched Business Hub on November 10, 2026, there is also first-party ChatGPT data to compare against. Our launch guide to ChatGPT Business Hub explains how to claim the profile and see it.

Treat the audit as a fixed cost and the program that follows as the variable one. A good audit is cheap relative to a year of content and PR, and it decides where that money goes.

Include every assistant your buyers use, not only ChatGPT. Similarweb, a traffic analytics vendor, reports that ChatGPT’s share of worldwide generative AI web traffic fell from roughly 76% to about 53% in a year, and that Gemini’s app reached about 716 million monthly users in May 2026. Forrester found business buyers are four times as likely as consumers to use Microsoft Copilot, often in company versions.

Ask sales which assistants buyers mention in calls and which tools your largest customers use internally. If most target accounts run Microsoft 365, Copilot deserves the same weight as ChatGPT in the audit, even though its consumer audience is smaller. If buyers are technical, Perplexity often matters more than its traffic share suggests, because it shows its sources and is used for research.

1. Build the question set: 30 to 50 questions buyers ask before they shortlist, grouped by stage (problem, category, comparison, pricing, alternatives, implementation). Use sales call notes and search queries, not your own product language.
2. Pick three to five competitors: the ones you lose deals to, not the biggest names in the market.
3. Run each question in each engine, in a clean session, at least twice, and save the full answer with the date.
4. Score each answer on five fields: named (yes or no), position in any list, accuracy of what is said, sources cited, and competitors named.
5. Add first-party data: Business Hub mentions and clicks for ChatGPT and Bing AI Performance citations for Copilot.
6. Summarise by stage and engine, then list the ten questions and five third-party sources with the biggest gaps.

Answers vary from run to run, so treat a single run as a sample. Two runs per question per engine, repeated monthly, give a usable trend. Our guide to measuring share of answer in ChatGPT, Perplexity and AI Overviews explains the method in more depth.

• Absent everywhere on category questions: your site and listings do not describe the category clearly. Start with answer-first category pages and directory listings.
• Named on brand questions, absent on comparison questions: competitors own the comparison content. Publish your own comparisons and alternatives pages.
• Named but described wrongly: trace each error to its source and fix it there. Our guide to fixing wrong AI answers about your company sets out the steps.
• Strong in one engine, weak in another: check crawler access and the sources each engine cites; they differ.
• Competitors cited from the same few sites: those sites are your PR and review targets for the next two quarters.

The audit should end with a decision, not a deck. Group the gaps into the five budget lines most AI visibility programs need: measurement, answer-first content, third-party authority, technical and data, and profile governance. Then estimate effort for each gap and rank by the value of the question, using deal data where you have it.

In our view, three patterns justify investment straight away: competitors named on comparison and alternatives questions you lose deals on, material errors about products or pricing, and a small number of third-party sites cited in most answers where you are missing. If the audit shows you already lead on most priority questions, the budget should go to protecting the position and to conversion, not to more content.

For a quick first view, our free AI visibility checker shows how ChatGPT, Perplexity, Gemini and Google AI Overviews describe you against competitors. A full audit adds your own question set, scoring across engines, first-party data and a prioritised backlog.

Our generative engine optimization team can run a full audit for your company and turn it into a 12-month plan.

If you want to see where you stand before you commit budget, talk to our team and we will review your category in a free 30-minute pipeline plan.