How do I compare AI visibility by topic cluster?

To compare AI visibility by topic cluster, the key is to measure how often your brand appears in AI answers across a set of related keywords/pages, then roll that up into clusters so you can compare performance by theme.

1) Group keywords into topic clusters

Start by grouping your keywords, pages, or prompts into clusters like:

• Product category
• Use case
• Industry
• Problem/solution
• Competitor comparisons
• Brand terms

Example:

• CRM for startups
  • best CRM for startups
  • startup CRM tools
  • CRM for small teams
• Email marketing
  • email marketing platform
  • best email automation software
  • newsletter tools

Each cluster should represent one user intent/theme.

2) Track AI visibility per query

For each keyword or prompt, check whether your brand appears in AI-generated responses from tools like:

• ChatGPT
• Google AI Overviews
• Perplexity
• Claude
• Bing Copilot

For each query, record:

• Mentioned or not
• Position/order if applicable
• Citation/source inclusion
• Sentiment/context
• Competitors mentioned
• Which pages are cited

3) Calculate visibility metrics by cluster

Once you have query-level data, aggregate it by cluster. Useful metrics:

A. Mention rate

% of prompts in the cluster where your brand is mentioned.

Formula: Mentions / Total prompts in cluster

B. Citation rate

% of prompts where your site/content is cited as a source.

C. Share of voice

How often your brand appears compared with competitors in the same cluster.

Formula: Your mentions / Total competitor mentions in cluster

D. Average rank/order

If the AI lists multiple brands, calculate your average position.

E. Coverage

How many queries in the cluster you appear in at all.

4) Compare clusters side by side

Create a table like this:

Topic Cluster	Queries	Mention Rate	Citation Rate	Avg Position	Top Competitors
CRM for startups	12	67%	42%	2.3	HubSpot, Pipedrive
Email marketing	10	20%	10%	4.1	Mailchimp, Klaviyo
Sales automation	8	50%	25%	3.0	Outreach, Apollo

This shows which themes are strong vs weak.

5) Normalize for cluster size

Don’t compare raw mention counts alone. A cluster with 30 prompts will naturally have more mentions than one with 5. Use rates/percentages so clusters are comparable.

6) Separate AI surfaces if possible

Visibility can differ by platform. Compare clusters across:

• AI Overviews
• ChatGPT
• Perplexity
• Copilot

You may find one cluster performs well in one surface but poorly in another.

7) Add quality/context scoring

A simple mention isn’t always valuable. Score whether the mention is:

• Positive
• Neutral
• Negative
• In a recommendation list
• In a comparison against a competitor
• Cited with your own content

That gives you a more useful “quality-adjusted visibility” metric.

8) Use the results to prioritize content

If a cluster has low visibility, it may need:

• Better topical coverage
• Stronger entity signals
• More FAQs and comparison pages
• More citations/authority
• Improved schema and internal linking

---

Simple workflow

1. Build topic clusters
2. Run a set of AI prompts for each cluster
3. Record mentions/citations/position
4. Aggregate into cluster-level metrics
5. Compare against competitors
6. Prioritize content updates based on gaps

If you want, I can also give you:

• a spreadsheet template
• a formula set for scoring AI visibility
• or a step-by-step method for ChatGPT/Perplexity/AI Overviews specifically.