How to Measure AI Visibility and Brand Mentions

To measure AI visibility, build a fixed list of questions your customers ask, run them regularly across the main AI assistants, and record whether your brand is mentioned, where it appears, how it is described and which sources are cited. Tracked over time, these numbers show whether your visibility is growing. There is no official AI ranking report, so you have to build this view yourself or use a tracking tool.

Step 1: Build a prompt set that reflects real buyers

Start with 30 to 100 prompts split into groups:

• Category prompts: “best accounting firms for startups in Paris”
• Problem prompts: “how do I reduce restaurant food waste?”
• Comparison prompts: “Brand A vs Brand B for small teams”
• Brand prompts: “what does [your brand] cost?” or “is [your brand] reliable?”

Use the language customers use. Sales calls, support tickets and site search logs are good sources. Keep the set stable so results are comparable month to month.

Step 2: Choose the assistants to track

Most brands track ChatGPT, Gemini, Google AI Overviews, Perplexity, Copilot and Claude. Answers vary by model, by whether web search is on, and by location, so note the settings you use each time.

Step 3: Record the right metrics

1. Mention rate: the share of prompts where your brand appears.
2. Position: first named, in the middle, or last.
3. Share of voice: your mentions compared with each competitor’s.
4. Accuracy: are prices, location, products and claims correct?
5. Sentiment: is the description positive, neutral or negative?
6. Citations: which URLs the assistant links to, including your own.

Our page on brand mention frequency in AI answers explains how to read these figures.

Step 4: Run several samples

AI answers are not fixed. The same prompt can give different brands on different runs. Run each prompt more than once, or across several days, and look at averages rather than single answers. A brand mentioned in three out of five runs is in a different position from one mentioned once in five.

Step 5: Connect visibility to traffic

Some AI tools send visits you can see in analytics. Look for referrals from domains such as chatgpt.com, perplexity.ai, copilot.microsoft.com and gemini.google.com. The numbers are often small and incomplete, because many users read the answer without clicking, so treat them as one signal among several. Also ask new customers “how did you hear about us?” and add “an AI assistant” as an option.

Step 6: Turn data into actions

Measurement is only useful if it leads somewhere. Typical findings and responses:

• A competitor appears for comparison prompts and you do not: publish an honest comparison page and seek coverage in review sites.
• An assistant gives wrong pricing: update your pricing page and profiles.
• You are mentioned but never cited: create a clear source page worth linking to.

For ideas on fixing inaccurate answers, see how to control how AI describes your brand.

Common measurement mistakes

Avoid changing your prompt list every month, because you lose the ability to compare. Do not judge results from one assistant only; a brand can be strong in Perplexity and missing in Gemini. And do not count a mention as a win if the description is wrong or negative. Accuracy and tone matter as much as frequency.

How Keboost helps

Doing this by hand takes hours each month. Keboost tracks the prompts your customers ask across AI models, shows where your brand and competitors appear, and highlights changes over time. It then creates and distributes content aimed at the gaps it finds. Run a free AI visibility check to get a first snapshot, or see plans for ongoing tracking.