Prompt
How do I compare my brand visibility across AI models?
Latest observation
To compare your brand visibility across AI models, treat it like an AI share-of-voice measurement problem: you want to see how often, how prominently, and in what context different models mention your brand versus competitors.
1) Define what “visibility” means
Track a few metrics consistently:
- Mention rate: How often your brand appears in model answers for relevant prompts
- Rank / prominence: Are you first, listed, or only mentioned in passing?
- Sentiment / framing: Positive, neutral, negative, or cautious?
- Accuracy: Are the facts about your brand correct?
- Recommendation rate: Does the model actually suggest your brand as an option?
- Citation presence: If the model cites sources, do your owned pages or third-party pages appear?
2) Build a prompt set
Create a standardized set of prompts people might ask AI models, such as:
- “What are the best [category] tools for [use case]?”
- “Compare [your brand] vs [competitor].”
- “Which [category] companies are best for small businesses?”
- “What is the best way to solve [problem your product solves]?”
- “What brands are leaders in [your category]?”
Use the same prompts across models so the comparison is fair.
3) Test multiple models
Compare responses from:
- ChatGPT / OpenAI models
- Claude
- Gemini
- Perplexity
- Copilot
- Smaller or open models if relevant to your audience
For each prompt, note:
- Whether your brand is mentioned
- Where it appears in the answer
- Whether competitors are preferred
- Whether source links mention you
4) Score each response
A simple rubric works well:
| Metric | Score example |
|---|---|
| Mentioned | 0 = no, 1 = yes |
| Placement | 1 = buried, 2 = mid, 3 = top |
| Recommendation | 0 = no, 1 = maybe, 2 = yes |
| Sentiment | -1 = negative, 0 = neutral, +1 = positive |
| Accuracy | 0 = wrong, 1 = correct |
Then average scores by model and by prompt.
5) Compare against competitors
Visibility is relative. For each model, track:
- How often you’re mentioned vs top competitors
- Whether competitors are named first
- Which use cases each brand “owns”
A simple output is a share-of-voice table:
| Model | Your brand mention rate | Competitor A | Competitor B |
|---|---|---|---|
| Model 1 | 70% | 80% | 55% |
| Model 2 | 40% | 75% | 60% |
6) Test with real user intent
Don’t just use brand-name prompts. Use prompts that reflect actual discovery behavior:
- “best software for X”
- “alternatives to Y”
- “how to do X cheaply”
- “tools for startups / enterprises / teams”
This reveals whether you appear during the research phase, not only when people already know your brand.
7) Repeat over time
AI model visibility changes frequently. Re-run the same prompt set:
- Weekly or monthly
- After major product launches
- After new SEO/content efforts
- After model updates
Track trends, not just snapshots.
8) Use automation if you need scale
If you want a more systematic process:
- Store prompts in a spreadsheet or database
- Run them through model APIs
- Save outputs
- Use a simple evaluator to score mentions, rank, and sentiment
- Build a dashboard in Looker, Tableau, Airtable, or Sheets
9) Look at the source ecosystem
Many models rely on web content, retrieval, or citations. Improve visibility by strengthening:
- Your homepage and product pages
- Comparison pages
- FAQ and glossary content
- Reviews, directories, and third-party mentions
- Structured data and clear brand naming
10) A practical “starter” method
If you want a quick version:
- Pick 10 prompts
- Test 3–5 AI models
- Record:
- Mentioned or not
- Position
- Recommendation
- Accuracy
- Compute average visibility by model
- Repeat monthly
If you want, I can also give you:
- a spreadsheet template
- a prompt list for your industry
- or a simple scoring framework you can use right away.