Quick answer
An AI visibility audit checks whether AI platforms name, describe and cite your brand accurately. Work through setup, prompts, platforms, measurement, gaps and actions. Record sample sizes, keep full answers and verify changes afterwards.
Key takeaways
- Set up before you measure.
- Keep the full answers.
- Choose a few actions and verify them.
Setup
0 of 4Measurement
0 of 4Analysis
0 of 4Action
0 of 3How to use this checklist
Work through the sections in order. Setup comes first because everything downstream depends on it. Keep the checklist with the report so that anyone reading the results can see what was and was not done. A skipped step is not a failure, provided it is written down.
Setup: getting the foundations right
- Confirm the brand facts in writing: what you do, who it is for, plans, prices, locations and differences.
- Confirm the domains and profiles that belong to the brand, so results can be attributed correctly.
- Agree the competitor set and the reason each one is included.
- Record who approves changes to the site and who owns each fix.
Measurement: a method you can defend
- Write the prompt set with topics, intents and prompts kept separate.
- Decide the platforms, the locations and the languages, and keep them fixed for the audit.
- Run every prompt several times per platform and store full answers with their sources.
- Exclude failed runs from denominators instead of counting them as misses.
- Record the sample size next to every rate.
Analysis: from numbers to causes
- Split every rate by platform, topic and intent before combining.
- Read a sample of answers behind each surprising number.
- List omissions, source gaps and wrong or outdated claims separately.
- Rank findings by frequency and by the damage they could do.
Action and verification
- Choose a small number of actions, each with an owner and a measure.
- Record a baseline before each change goes live.
- Schedule a rerun of the same prompts, and repeat it more than once.
- Report each action as verified, inconclusive or not yet observable.
Common ways audits go wrong
After the audit: turn it into a habit
An audit is most valuable as the first cycle of a routine, not a one-off report. Keep the prompt set, the platform list and the definitions in a place the whole team can see, and record the date each part was last reviewed. When the next cycle starts, you should be able to repeat the run exactly, which is what makes a comparison between cycles meaningful.
Share the result in two forms: a short summary for people who need the headline, and the full evidence for people who want to check it. The summary earns attention and the evidence earns trust. If someone challenges a number, you should be able to open the answers behind it within a minute.
How long does an audit take?
It depends on scope: brands, prompts and platforms. Agree the scope first, then plan time for review as well as collection.
How often should I repeat the audit?
A light monthly rerun of the same prompt set suits most teams, with a fuller review each quarter. Rerun sooner after any significant change.
Cite this page
BrandWater AI Research. (2026, 19 September 2026). The AI visibility audit checklist. https://brandwaterai.in/guides/ai-visibility-audit-checklist
How we work
Figures are dated and linked to their sources. Where none exist we say so. Read our methodology and AI transparency pages.