← All postsAI Search

How to Track Whether ChatGPT is Recommending Your Business

A practical guide to monitoring whether AI models like ChatGPT, Claude, and Perplexity are recommending your business, and how to track changes over time.

5 min read
How to track whether ChatGPT is recommending your business

You Cannot Optimise What You Do Not Measure

If you care about AI visibility, you need to know where you currently stand. Unlike Google rankings, which you can check with a simple search, AI citation is harder to track because AI responses are conversational, context-dependent, and vary between users.

Here is a practical framework for monitoring your AI visibility across ChatGPT, Claude, and Perplexity.

Step 1: Define Your Target Queries

Start by listing the queries potential customers might ask AI about your business category. For a dental clinic in Orchard Road, these might include "recommend a dentist near Orchard Road," "best dental clinic Singapore," "dentist for wisdom tooth extraction Singapore," and "affordable teeth whitening Orchard."

Be specific. Include location, service type, and common modifiers that customers use. Aim for 10 to 20 target queries that represent your most valuable customer searches.

Step 2: Test Each Query Across Models

Open ChatGPT, Claude, and Perplexity in separate tabs. Ask each model your target queries and record whether your business is mentioned. Note the exact response: are you recommended by name? Mentioned as one of several options? Not mentioned at all?

Important: AI responses are not static. Test from a fresh conversation (not a thread where you have previously discussed your business). Test at different times. Test with slight variations of the query. This gives you a more accurate picture of your baseline visibility.

Step 3: Score Your Visibility

Create a simple tracking sheet with your target queries as rows and the AI models as columns. For each combination, score your visibility. A score of 3 means you are recommended by name with positive framing. A score of 2 means you are mentioned among several options. A score of 1 means your category is discussed but you are not named. A score of 0 means no mention at all.

Average your scores across all queries and models. This is your baseline AI visibility score.

Step 4: Track Monthly Changes

Repeat this assessment monthly. AI models update their knowledge bases, adjust their algorithms, and incorporate new training data. Your visibility can change over time, both positively (as you improve your AI presence) and negatively (as competitors improve theirs).

Track the trends. Is your visibility increasing, stable, or declining? Which queries show the most improvement? Which models cite you most frequently?

Step 5: Monitor Accuracy

When AI models do mention your business, check whether the information is accurate. Incorrect pricing, outdated service descriptions, wrong addresses, and other errors can harm your business even when you are being cited.

If you find errors, this is a signal that your structured data and web presence need updating. AI models can only be as accurate as the information available to them.

Automated Monitoring

Manual tracking is useful for understanding your current position but does not scale. The Swop Labs AI Search team provides automated monitoring as part of our AEO engagements. We track citation frequency, accuracy, competitive positioning, and changes over time across all major AI models.

For businesses handling their own monitoring, consider running your test queries monthly and maintaining a spreadsheet that tracks changes. Even this basic approach gives you valuable insight into whether your AEO efforts are working.

What to Do With the Data

If your visibility score is low (under 1.0 average), focus on the foundational work: structured data, content authority, and entity recognition. Our AEO guide covers these.

If your visibility is moderate (1.0 to 2.0), focus on differentiators: what makes your business the one AI should recommend over competitors? Strengthen your review profile, build content authority, and consider implementing an llms.txt file or MCP server.

If your visibility is high (above 2.0), focus on maintaining and expanding. Monitor for accuracy, track competitors entering your space, and ensure your structured data stays current.

Talk to Us

Chat with us on WhatsApp to discuss AI visibility monitoring for your business. We reply within one Singapore business day.

Ready to get started?

Chat with the Swop Labs team on WhatsApp. We reply within one Singapore business day.

Chat on WhatsApp

More from the blog