How do you measure visibility in AI search engines?

Unlike SEO, where Google Search Console gives you your rankings in real time, measuring GEO visibility calls for a different approach. There’s still no standard tool yet, but reliable methods do exist. Sandro, cofounder of the Gemeos Webflow agency, walks you through the full process.

Why GEO measurement is different from SEO

In SEO, you measure your position for a given keyword. In GEO, you measure whether your brand, website, or content is mentioned in an AI-generated answer. The variables are different: each LLM has its own sources, answers change depending on how the question is phrased, and there’s no fixed “position 1.”

1. Build your target query list

Before you measure anything, define the 10 to 20 queries you want to appear for in AI answers. Cover three categories:

• Brand queries: “what is [your company],” “reviews of [your brand]”
• Service queries: “best Webflow agency in France,” “what tool should I use to do X”
• Industry queries: “how to do X in [your field],” “definition of Y”

2. Monthly manual testing protocol

Once a month, test each query on your list in the three main AI engines. Log the results in a tracking sheet.

1. Open a private browsing session to avoid personalization
2. Test each query in ChatGPT (GPT-4), Perplexity, and Gemini
3. Note whether your brand or URL is mentioned, and in what context
4. Note which competitors are mentioned instead of you
5. Test 2 to 3 different phrasings of the same query

3. Available GEO tracking tools

Several emerging tools can automate this tracking. None are as mature as traditional SEO tools yet, but they’re getting more reliable:

• Profound: brand mention tracking in LLM answers, with alerts
• Otterly.ai: monitoring citations in Perplexity and ChatGPT for defined queries
• Brandwatch AI Insights: tracking brand mentions in AI-generated content
• Mention: broader monitoring including emerging AI platforms

4. GEO metrics to track

5. Create your GEO dashboard

Use a Google Sheet or Airtable to track your monthly results. Recommended columns: test date, query, AI engine, cited (yes/no), position in the answer, competitor cited instead, notes. This manual tracking takes 30 to 60 minutes a month and gives you a clear view of your progress.

Conclusion

GEO measurement is still empirical, but setting up a testing protocol now lets you build a baseline and measure the impact of your optimizations over time.

• Use case 1: an agency that wants to show clients the impact of its GEO work with concrete data
• Use case 2: a startup that wants to monitor whether competitors are mentioned more often than it is for its target queries
• Use case 3: a consultant who wants to prove visibility in their field of expertise in AI answers