Sample report

What a Pharma AI Monitor readout looks like.

A fictional in-market brand, ExampleRx, tracked across ChatGPT, Gemini, Perplexity, and Grok. See the brand view, three findings, and the action plan.

ExampleRx

In-market · Fictional indication · US

Illustrative workspace · Sample data

BRAND OVERVIEW6 Sep 2026 → 20 Sep 2026 · Illustrative cycles

ExampleRx is still named in broad answers. It is losing the switching conversation.

• Share of Answer fell from 80% to 72.5%.
• 100% of the drop is in “after first-line treatment” questions.
• RivalOne and RivalTwo are named instead.

Open the switching gap

AI Pulse score74▼4

Share of Answer · ExampleRx

72.5%7.5 pts

348 of 480 answers
Previously 384 of 480 (80%)

% of answers naming the brand in this question set. Not patient reach.

Scope: 40 questions × 4 AI platforms × 3 repeats = 480 answers per cycle. Same questions both cycles. All data is fictional.

Answers naming ExampleRx, by question group

6 Sep20 Sep

Broad treatment options108 → 108 / 120

After first-line treatment108 → 72 / 120

Administration & dosing84 → 84 / 120

Long-term treatment84 → 84 / 120

Each group: 10 questions × 4 platforms × 3 repeats. Two cycles show a change, not a trend.

Three findings

Open a finding to see the answer behind it.

Each shows the question, the full AI answer, cited sources, and what to do next.

AI Pulse is the headline. Each finding stands on its own evidence.

Visibility

“Which options could an HCP consider after first-line treatment has not worked in this indication?”

Record: EX-017 · repeat 2 of 3Platform: ChatGPTCaptured: 20 Sep 2026 · 10:00 UTCFictional answer, no live model run

Full AI answer

An HCP might consider RivalOne or RivalTwo after reviewing the patient's response to first-line treatment. The choice depends on the patient's clinical circumstances and the approved prescribing information. The treatment options overview [1] and patient discussion guide [2] describe what to consider.

Cited sources

[1] Treatment options overviewclinical-review.example/treatment-options · Fictional source

Example Clinical Review

Discusses RivalOne and RivalTwo for patients switching treatment. Does not mention ExampleRx.

[2] Discussing a treatment changepatient-guide.example/treatment-change · Fictional source

Example Patient Guide

Suggests patients discuss response to prior treatment with their doctor. Does not list ExampleRx.

What we see

ExampleRx is not named. RivalOne and RivalTwo are. This question group accounts for all 36 lost mentions.

So what

ExampleRx is losing the switching conversation, not the category. That is a specific, fixable gap.

Next steps · Brand + Omnichannel

Pull all 48 switching answers that omit ExampleRx.
List the pages they cite. Both top sources skip ExampleRx.
Check whether your HCP site answers the switching question directly.
Brief the agency on an MLR-approved switching page.

One answer shows the gap. It does not prove why the model left ExampleRx out.

Visibility

“Which options could an HCP consider after first-line treatment has not worked in this indication?”

Record: EX-017 · repeat 2 of 3Platform: ChatGPTCaptured: 20 Sep 2026 · 10:00 UTCFictional answer, no live model run

Full AI answer

Cited sources

[1] Treatment options overviewclinical-review.example/treatment-options · Fictional source

Example Clinical Review

Discusses RivalOne and RivalTwo for patients switching treatment. Does not mention ExampleRx.

[2] Discussing a treatment changepatient-guide.example/treatment-change · Fictional source

Example Patient Guide

Suggests patients discuss response to prior treatment with their doctor. Does not list ExampleRx.

What we see

ExampleRx is not named. RivalOne and RivalTwo are. This question group accounts for all 36 lost mentions.

So what

ExampleRx is losing the switching conversation, not the category. That is a specific, fixable gap.

Next steps · Brand + Omnichannel

Pull all 48 switching answers that omit ExampleRx.
List the pages they cite. Both top sources skip ExampleRx.
Check whether your HCP site answers the switching question directly.
Brief the agency on an MLR-approved switching page.

One answer shows the gap. It does not prove why the model left ExampleRx out.

Message credit

“How do the dosing schedules of ExampleRx, RivalOne, and RivalTwo compare?”

Record: EX-026 · repeat 1 of 3Platform: GeminiCaptured: 20 Sep 2026 · 10:05 UTCFictional answer, no live model run

Full AI answer

The options differ in how they are given. RivalOne is described as offering a less frequent schedule. ExampleRx and RivalTwo are also discussed, but their schedules are not compared. The administration comparison [1] gives an overview; check current prescribing information for dosing details.

Cited sources

[1] Comparing administration schedulestreatment-review.example/administration · Fictional source

Example Treatment Review

Credits RivalOne with less frequent dosing. Describes ExampleRx without mentioning its schedule.

What we see

The dosing benefit appears in the answer, credited to RivalOne. ExampleRx is named but gets no credit.

So what

Being named is not the same as owning the message. The cited comparison page frames the benefit around RivalOne.

Next steps · Brand + Agency + Medical

Confirm the approved dosing message and its PI support.
Check how clearly your owned pages state the schedule.
Review how the cited comparison page describes each brand.
Brief the agency on any owned-content gap.

This does not claim ExampleRx has the same benefit, or that the competitor's description is wrong.

PI review

“What does the evidence say about long-term benefit with ExampleRx?”

Record: EX-034 · repeat 3 of 3Platform: PerplexityCaptured: 20 Sep 2026 · 10:10 UTCFictional answer, no live model run

Full AI answer

ExampleRx provides lasting benefit for all patients. A long-term treatment overview [1] describes sustained outcomes. Patients should discuss treatment expectations with their healthcare professional and consult the current prescribing information.

Cited sources

[1] Long-term treatment overviewhealth-summary.example/longer-term · Fictional source

Example Health Summary

Uses broad language about “sustained benefit” without stating the study population or follow-up period.

PI comparison · Needs human review

Answer claim: “ExampleRx provides lasting benefit for all patients.”
Fictional PI reference · v1 · §14: “Benefit was evaluated in the study population during the specified follow-up period.”

Possible overstatement of population and duration. Needs human review.

The PI is the comparison standard. It is separate from the sources the AI cited.

What we see

The answer says “lasting benefit for all patients.” The PI limits benefit to the study population and follow-up period.

So what

The answer overstates population and duration. Medical should decide whether it needs escalation.

Next steps · Medical / MLR

Compare the full statement with PI §14.
Check the cited page for the source of the wording.
Decide: escalate, contact the publisher, or log and monitor.
Record the decision with the answer attached.

A fictional review flag, not a clinical assessment of a real therapy.

Action plan

Who does what next.

Owner · Brand + Omnichannel

Missing from switching answers

Pull all 48 switching answers that omit ExampleRx.
List the pages they cite. Both top sources skip ExampleRx.
Check whether your HCP site answers the switching question directly.
Brief the agency on an MLR-approved switching page.

Re-run: Re-run the switching questions next cycle. Log content changes separately.

Owner · Brand + Agency + Medical

Competitor credited with your dosing benefit

Confirm the approved dosing message and its PI support.
Check how clearly your owned pages state the schedule.
Review how the cited comparison page describes each brand.
Brief the agency on any owned-content gap.

Re-run: Compare message credit on the same dosing questions next cycle.

Owner · Medical / MLR

AI overstates long-term benefit

Compare the full statement with PI §14.
Check the cited page for the source of the wording.
Decide: escalate, contact the publisher, or log and monitor.
Record the decision with the answer attached.

Re-run: Re-check this question and related wording next cycle.

Your teams set priorities and approve all content changes.

Fictional data. No live AI runs or customer data.How we measure

What would this show for your brand?

We agree your indication, competitors, and questions, then walk your team through the baseline.

Request a baseline