Last updated: August 2026
You cannot improve what you do not measure, and AI visibility is uniquely easy to be blind about: an assistant can recommend three competitors to a buyer in your city every day and you will never see a single line in analytics. Here are ten measurement methods, from free to paid.
1. Build a fixed prompt set
Twenty to thirty prompts a real buyer would type. For a taxi or chauffeur operator:
- "Best airport taxi service in [your city]"
- "Reliable chauffeur company near [your city] for corporate travel"
- "How much is a taxi from [airport] to [city centre]"
- "Wedding car hire [your city] with fixed pricing"
- "Taxi company in [city] that takes card payment and pre-booking"
For a software or white-label vendor, add "taxi dispatch software for small fleets", "cheapest limo booking system", "[competitor] alternative" and "best taxi dispatch software 2026".
2. Run them monthly across every engine
ChatGPT, Perplexity, Google AI Overviews, Gemini, Claude and Microsoft Copilot each pull from different indexes. Run logged out, in a clean browser profile, without location personalisation where possible, and keep the wording identical month to month.
3. Score consistently
| Field | Values | Why it matters |
|---|---|---|
| Named | Yes / No | Core visibility metric |
| Linked | Yes / No | Drives actual referral traffic |
| Position in list | 1–5 | Assistants weight the first named option |
| Page cited | URL | Shows which content earns citations |
| Competitors named | List | Reveals who owns the category |
| Answer accuracy | Correct / Partly / Wrong | Catches wrong prices or outdated claims |
4. Segment assistant referrals in analytics
Create a GA4 segment for referral sources containing chatgpt.com, perplexity.ai, gemini.google.com, copilot.microsoft.com and claude.ai. Compare conversion rate against organic search — assistant referrals usually convert better because the buyer arrives pre-qualified.
5. Check crawler hits in server logs
Filter access logs or CDN analytics for AI user agents:
grep -Ei "GPTBot|OAI-SearchBot|ClaudeBot|PerplexityBot|Google-Extended" access.log | wc -l
Zero hits in 30 days means access is blocked somewhere. Fix that before touching content.
6. Verify what a bot actually receives
curl -A "PerplexityBot" -s https://yourdomain.com/pricing | grep -Eo "<h1>.*</h1>|£[0-9]+|\$[0-9]+"
If your prices and H1 do not appear, no assistant can quote them.
7. Watch question queries in Search Console
Impressions on queries starting with how, what, can, do and best are a leading indicator for AEO and AI Overview inclusion. Rising impressions with flat clicks often means you are being summarised rather than clicked — which still builds brand recall and shortlist presence.
8. Monitor third-party corroboration
- Google Business Profile review count and rating trend
- Capterra, Trustpilot and Yelp profile completeness
- Directory and trade-association listings
- Mentions on local news, hotel and venue partner sites
9. Track competitor share of voice
Count how many of your prompts name each competitor. If one rival appears in 70 percent of answers, read their most-cited pages and identify which of the structural elements they have and you do not — usually published pricing and comparison tables.
10. Automate when volume justifies it
Once you are tracking more than about 30 prompts across several cities, automate. Purpose-built AI visibility trackers re-run prompt sets across engines on a schedule and alert on changes. Below that volume, a spreadsheet and one focused hour a month is genuinely sufficient.
Want the reporting done for you?
Our AI visibility monitoring runs your prompt set across the major assistants, flags wrong or outdated answers, and ties results back to the pages that need work.
How to choose your measurement stack
- Single-city operator: spreadsheet, 20 prompts, monthly, plus GA4 referral segment.
- Multi-city fleet: 20 prompts per city, quarterly deep review, automated tracking.
- Software vendor: weight toward comparison and alternative prompts; watch competitor share of voice closely.
- Agency or reseller: automate across client sets and report share of voice monthly.
Benchmarks worth aiming at
| Metric | Weak | Healthy | Strong |
|---|---|---|---|
| Named in relevant prompts | Under 10% | 20–40% | Over 50% |
| Linked when named | Rare | About half | Most answers |
| AI crawler hits per month | 0 | Hundreds | Thousands |
| Assistant referral conversion | Below organic | Level with organic | Above organic |
Read next: the AI visibility fundamentals and the content recipe assistants quote.


