GuidesAI search
Tutorial: tracking your brand's visibility in AI answers
1. The question list
Thirty to fifty real questions from your customers, in three families:
| Family | Example |
|---|---|
| Generic | “Which agency should build my mobile app?” |
| Comparison | “Shopify or WooCommerce for a small store?” |
| Brand | “Is [your brand] any good?” |
2. The record
date,assistant,question,brand_mentioned,link_to_site,position,competitors_cited
2026-11,ChatGPT,"Which agency should…",yes,yes,2,"A;B"
2026-11,Perplexity,"Which agency should…",no,no,,"A;C"
3. The metrics
- Mention rateShare of questions where the brand appears.
- Citation rateShare of answers with a link to your site.
- CompetitorsWho is cited instead of you, and on which questions.
4. Cross-check with traffic
Visits from assistants show in GA4, and their bots’ visits in the server logs. The three measures together tell you whether your efforts are paying off.
Traffic: measuring ChatGPT traffic in GA4. Bots: spotting AI bots in your logs. To get cited: generative engine optimization.
Frequently asked questions
- Do answers change from one time to the next?
- Yes, even for the same question. That's why you track a rate across dozens of questions rather than a single answer.
- Are there tools for this?
- Several services automate the check. A spreadsheet and an hour a month are enough to start and understand what you're measuring.
- Does the signed-in account affect the answer?
- It can: history and memory personalise answers. Use a session with no history for the check.