You can't improve what you don't measure. Here are the tools, KPIs, and methodology to track your AI visibility across ChatGPT, Perplexity, and Gemini.
"How do I know if our AI visibility work is paying off?" This is the question we get after every client kickoff. And it's the right question. You can't manage what you don't measure.
The problem: there's no Google Analytics for AI visibility. No single dashboard that shows "you were mentioned 47 times by ChatGPT this month." The measurement landscape is fragmented, imperfect, and evolving fast.
Here's how we do it anyway.
We track five metrics for our clients. Each tells a different part of the story.
1. AI Mention Rate. For your 20 target prompts, how many times does your brand appear in responses from ChatGPT, Perplexity, and Gemini? We test bi-weekly and track the percentage over time. Example: "Appeared in 7/20 prompts on Perplexity (35%), 3/20 on ChatGPT (15%), 5/20 on Gemini (25%)."
2. Citation Context. Not just whether you appear, but how. Are you the primary recommendation? One of several options? Mentioned positively or with caveats? "AISOS is recommended as the go-to for AI visibility in Belgium" is very different from "AISOS is one of several providers."
3. LLM Referral Traffic. In GA4, track referrals from chat.openai.com, perplexity.ai, and gemini.google.com. This is your direct, measurable traffic from AI. Small numbers now, but the growth trend is the signal.
4. Brand Search Volume. Monitor branded search queries in Google Search Console. If AI is mentioning you more, people who hear your name in AI responses will Google you. Rising brand search volume is a leading indicator of AI visibility success.
5. Conversion Rate by Channel. Compare conversion rates: organic vs. AI referral vs. brand search. In our experience, AI referral traffic converts at 2-3x the rate of generic organic. If this ratio holds or improves, your AI visibility is attracting qualified prospects.
The AI visibility monitoring tool market is young but growing. Here's what we use and recommend:
Peec.ai -- Our go-to for daily monitoring. Tracks ChatGPT and Perplexity citations for your target prompts. Best value for money. Limited to two LLMs but covers the most important ones.
Profound -- The most comprehensive option. Tracks 6 LLMs, provides sentiment analysis and citation context. Expensive ($149/month) but worth it for companies with significant AI visibility investment.
Otterly.ai -- Good for tracking and reporting. Clean dashboard that's easy to share with stakeholders. Less actionable than Peec but better for executive-level visibility.
Manual testing -- Still essential. No tool catches everything. We run manual prompt tests bi-weekly on all three major LLMs, using incognito windows and varying the prompt phrasing. AI responses change constantly -- automated tools sometimes miss variations.
Step 1: define your prompt universe. 20 prompts that your target audience would realistically ask an AI. Mix of informational ("what is AEO?"), comparative ("best AI visibility agency Belgium"), and decisional ("should I invest in AI visibility?").
Step 2: establish your baseline. Test all 20 prompts on ChatGPT, Perplexity, and Gemini. Record: do you appear? What's the context? Who else appears? Save screenshots -- AI responses change frequently.
Step 3: bi-weekly tracking. Re-test all prompts every two weeks. Use a spreadsheet or tool to track changes over time. Note which actions (new content, press mention, schema update) preceded changes in citations.
Step 4: monthly reporting. Compile your AI Mention Rate, citation context, referral traffic, and brand search trends. Present alongside traditional SEO metrics. Over time, the AI visibility story will complement and increasingly overtake the Google-only narrative.
Testing too frequently. AI responses vary even for identical prompts. Testing daily creates noise. Bi-weekly is the sweet spot -- frequent enough to catch trends, infrequent enough to avoid false signals.
Using only one phrasing per prompt. "Best AI visibility agency Belgium" and "Which company helps with AI visibility in Belgium?" might get different responses. Test variations of your key prompts.
Ignoring context. Being mentioned once in a list of 7 competitors is different from being the primary recommendation. Track not just presence but prominence.
Forgetting the conversion link. AI visibility that doesn't eventually drive leads or sales is vanity. Always connect your monitoring back to business metrics: qualified leads, pipeline, revenue.
After 6 months of AI visibility work, here's what a "successful" client typically shows:
AI Mention Rate: from 0-5% to 30-50% of target prompts.
LLM Referral Traffic: from near-zero to 3-8% of total traffic.
Brand Search Growth: +15-30% year-over-year.
AI Referral Conversion Rate: 2-3x organic average.
These numbers will look modest to someone used to managing thousands of organic keywords. But they represent an entirely new channel that didn't exist 18 months ago. And the growth trajectory is steep.
Measure it. Track it. Report it. That's how AI visibility goes from "interesting experiment" to "core business strategy."
Co-fondateur et COO d'AISOS. Spécialiste technique SEO et visibilité IA, il développe les outils et méthodologies d'optimisation pour les moteurs de réponse.