TL;DR
Focus on four metrics: citation share (your % of AI responses in your category), citation context (recommendation vs neutral mention), referral conversion rate (do AI visitors buy?), and crawler coverage (what % of your pages do AI crawlers visit?).
Most AI metrics are vanity metrics
"We were mentioned by ChatGPT 47 times last month." Is that good? Compared to what? Without context, citation count is meaningless. It's the AI equivalent of tracking page views without tracking conversions.
The metrics that predict revenue are more nuanced. They measure competitive position, recommendation quality, and commercial impact — not just volume.
The four metrics that matter
1. Citation share
What percentage of AI responses in your category include your brand? If users ask 100 purchase-intent questions about your category and you appear in 15 responses, your citation share is 15%. Track this weekly. Compare against competitors. This is the AI equivalent of market share.
2. Citation context
Are you being recommended or just mentioned? A recommendation ("We suggest Brand X for...") has 10x the revenue impact of a neutral mention ("Brands in this space include...") and infinitely more than a negative citation. Classify each citation and track the ratio.
3. Referral conversion rate
When AI-referred visitors land on your site, do they buy? Segment AI referral traffic in your analytics and measure conversion rate vs other channels. In our data, AI-referred visitors convert 20-40% higher than organic search visitors because the AI already qualified the recommendation.
4. Crawler coverage
What percentage of your important pages are AI crawlers actually visiting? Check server logs. If GPTBot visits your blog but never your product pages, you're building authority without earning commercial citations. Crawler coverage tells you where the gap is.
Building a weekly dashboard
Track these four metrics weekly. Plot trends over 12 weeks. Look for leading indicators: citation share increases that precede revenue increases, crawler coverage drops that precede citation declines, context shifts from recommendation to neutral. These patterns let you predict revenue impact before it shows up in sales data.