How do I measure my website's performance in AI-powered search engines?
How to measure your website's performance in AI-powered search engines: citations, AI referral traffic, crawler access and conversions, with the tools for each.
Measuring a website in AI search is not one number. It is four separate questions, and each needs different data. Are AI engines mentioning and citing you? Can their crawlers reach your pages? Do any visitors arrive from them? And do those visitors do anything worth the effort? Most teams answer one of these and assume the rest.
Here is how to measure each, which free sources help, and where a dedicated platform earns its cost. Vendor details come from the catalog entries on this site (read on 24 October 2026) and Promptwatch's published documentation.
The four layers of AI search measurement
- Visibility: does your brand appear in answers for the prompts that matter, and how prominently?
- Citations: which of your pages and which third-party pages do the engines cite?
- Access: can AI crawlers fetch your pages, and what happens when they do?
- Outcomes: how many visits come from AI platforms, and how do they convert?
Classic SEO reporting covers parts of layers 3 and 4 only. Search Console shows Google traffic. GA4 shows sessions. Neither shows whether ChatGPT cited you, nor what a crawler such as GPTBot fetched.
What you can measure for free
GA4 and referral traffic
AI assistants send traffic that shows up as referrals. In GA4, look at the traffic acquisition report by session source and filter for domains such as chatgpt.com, perplexity.ai, gemini.google.com, copilot.microsoft.com and claude.ai. Create a custom channel group or an exploration that groups those sources as "AI". Two caveats: some AI visits arrive without a referrer and land in direct traffic, and clicks are a small slice of exposure, since most AI answers are read without a click. Treat GA4 as a floor, not a total.
Google Search Console
Search Console shows clicks and impressions from Google Search. Google's documentation says AI features in Search are reported within the Web search type, so you cannot split AI Overviews from classic results there. It is still the right place to confirm which queries and pages gain or lose impressions, and to check indexing. Bing Webmaster Tools gives equivalent data for Bing, which feeds Copilot.
Server logs
Your own server or CDN logs show which AI crawlers fetched which URLs. They are accurate, but raw logs take work to parse, and user agent strings can be spoofed.
Manual prompt checks
Run ten to twenty commercial prompts by hand each month in ChatGPT, Gemini, Perplexity and Google. Record who is named and what is cited. It is slow and the answers vary run to run, but it is a useful sanity check on any tool.
Where a platform helps
The free sources leave gaps: no systematic prompt tracking, no citation breakdown, no verified crawler view, and no join between the layers. That is the case for a dedicated tool.
1. Promptwatch: measures all four layers
Promptwatch ranks first because it covers each layer in one workspace:
- Visibility: Visibility Score, Share of Voice, Sentiment and Position in Answer, with daily prompt panels per model, country and language.
- Citations: page and domain level citations, Page Tracker for specific URLs, offsite mentions, Reddit, YouTube and LinkedIn reports, and Citation Trends with citation rate, first cited, peak rate and decay half-life per page.
- Access: Agent Analytics, real-time and IP-verified AI crawler logs with a citation rate per page, error tracking and failure explanations.
- Outcomes: visitor analytics that attribute AI-referred traffic and conversions from ChatGPT, Perplexity, Gemini, Claude, Copilot, DeepSeek and more, via a lightweight script, GTM template or WordPress plugin.
- The join: 360 Insights puts crawler logs, citation clicks and AI-referred visits side by side per page, and Search Insights overlays Google Search Console and Bing Webmaster traffic on AI crawls and citations.
TreeHouse, which runs Pararius and Huurwoningen.nl, published a case study stating that AI search became about 25% of its organic conversions, nearly all from ChatGPT, after fixing crawler access. That is TreeHouse's own result.
Pricing: Explore is free (10 prompts, ChatGPT only), Essential $95 a month with visitor analytics, Professional $245 with crawler logs, Business $579, and agency plans $199, $399 and $799. Crawler logs start on Professional.
2. AthenaHQ
AthenaHQ tracks up to 11 models on Starter ($295 a month, credit-based) and lists GA4, Search Console, Shopify and Webflow connections for attribution. The catalog notes no crawler-log product.
3. Scrunch AI
Scrunch AI offers Core at $250 a month with four engines and reports AI agent and bot traffic. Weekly refreshes are noted in the catalog.
4. Profound
Profound lists Answer Engine Insights and Agent Analytics on its Enterprise offering, with contracts estimated at around $40k a year.
5. SEO suites
Ahrefs Brand Radar adds AI crawler activity reporting and tracks Overviews, AI Mode, ChatGPT, Perplexity, Gemini and Copilot, from $199 a month for the Index. Semrush AI Toolkit and SE Ranking bolt AI tracking onto rank tracking you may already use.
6. Trackers
LLM Pulse lists AI traffic analytics from EUR 49 a month. Otterly.AI comes last for this use, since the catalog describes it as monitoring only with no traffic or revenue attribution.
Comparison table
| Tool | Visibility tracking | Citation detail | Crawler logs | AI-referred conversions |
|---|---|---|---|---|
| Promptwatch | Daily, all active models | Page, domain, Reddit, YouTube, LinkedIn | Yes, from Professional | Yes |
| AthenaHQ | Up to 11 models | Citation Engine on Enterprise | Not listed | GA4 attribution listed |
| Scrunch AI | 4 engines on Core | Not documented in depth | Bot traffic reporting | Not listed |
| Ahrefs Brand Radar | Index plus Custom Prompts | Sources, YouTube, Reddit | Crawler activity reporting | Not listed |
| Otterly.AI | 4 base engines | Citation source tracking | Not listed | No |
Who should pick which
- You want all four layers in one place: Promptwatch.
- You are testing the idea: GA4 referral filter, Search Console and manual checks first, plus the free Promptwatch AI Visibility Report.
- You live in an SEO suite: try its add-on, and add a crawler-log source separately.
A simple monthly scorecard
Track five numbers each month: share of tracked prompts where you are named, number of cited pages, AI crawler hits and errors on key pages, AI-referred sessions, and AI-referred conversions. Trend them, not a single month's value. If visibility rises but conversions do not, look at which pages are cited and what they say. If crawler errors are high, fix access first. No metric guarantees a recommendation, but the scorecard tells you what to work on next.