How we measure your brand in AI answers.
A fixed question set, run across four AI platforms every cycle, with every answer saved. Here is what we count, and what the numbers do and do not mean.
Four steps, repeated every cycle.
- 01
Scope
Agree the brand, indication, market, competitors, key messages, and patient and HCP questions. Pre-launch scope covers disease and mechanism only.
- 02
Capture
Run the questions across ChatGPT, Gemini, Perplexity, and Grok, with repeat runs. Save the full answer, platform, model where available, time, and cited links.
- 03
Analyze
Measure brand mentions, positioning, message credit, and possible PI conflicts. Compare like-for-like runs only.
- 04
Re-run
Run the same questions each cycle. Any change in scope is shown in the comparison.
A number is only as good as its scope.
Visibility measures presence in the answers we collected. It is not patient reach or prescription share.
Know what is counted
Share of Answer = answers naming the brand ÷ all eligible answers. It is not patient reach or market share.
Compare the same scope
Same questions, platforms, market, and repeat runs. A changed question set changes the result.
Treat movement as a lead
AI answers vary between runs, and platforms update their models. Repeat runs show what is noise and what is real.
A flag starts a review. It does not end one.
We compare AI claims about your brand with your approved Prescribing Information (PI). Your reviewers make the call.
For Medical & MLREach flag shows
- The exact claim and the question that prompted it.
- Whether it contradicts, goes beyond, or omits PI context.
- The PI document and version used.
- The cited sources, as places to look.
A claim missing from the PI is not automatically false. Medical, Legal, and Regulatory teams make the final decision.
What each number means.
- AI Pulse score
- One headline score for how AI represents your brand, built from visibility, positioning, and PI alignment.
- Share of Answer
- % of answers in the question set that name your brand. Not patient reach or market share.
- Message credit
- Whether AI ties a key benefit to your brand, a competitor, or no one.
- PI flag
- An AI claim that may conflict with your Prescribing Information. A prompt for Medical review, not a verdict.
- Cited source
- A page the AI linked to in its answer. Shows where to look, not what caused the answer.
A defined benchmark
We monitor a fixed question set, not real patient conversations. No patient data is needed.
Citations show links, not causes
A cited page is where to look. It does not prove where a claim came from.
Lifecycle changes the read
Pre-launch: disease and mechanism understanding. In-market: brand visibility and messages.
See the method in action.
Follow one fictional finding from the brand view to the answer, its sources, and the next step.