Method notes
Clear definitions for providers, models, prompts, repeats, and scoring.
Research & field notes
Clear methods. No AI hype.
Practical guidance for pharma teams learning how to monitor and assess AI answers.
Field note 01
What is GEO for Pharma?
A practical introduction to how generative AI changes brand discovery and representation.
Field note 02
How to Measure Share-of-Answer
The definition, calculation choices, and limits behind a common AI visibility measure.
Field note 03
PI-Backed Claims: Why LLMs Hallucinate About Pharma Products
Why answer review needs source and claim context in a regulated category.
Field note 04
LLM Monitoring vs SEO vs Social Listening
Where these disciplines overlap, and where AI answers need a different approach.
Research program
What we should publish next
Benchmarks can build trust, but only when the method can hold up to scrutiny.
Audited benchmark
A labeled cohort with the dates, exclusions, and uncertainty shown.
Evidence patterns
Anonymized answer and source patterns that help regulated teams learn.
Team playbooks
Practical review routines for brand, insights, medical, and portfolio teams.
Important labeling rule
Samples are samples. Illustrative workflows are not customer case studies.
Until a named customer approves a result for publication, the redesigned experience labels fictional examples and composite workflows clearly. This preserves the useful stories from the current site without presenting them as verified customer outcomes.
Want to help shape the first auditable benchmark?
Start with a focused brand or portfolio baseline. Cohort participation, anonymization, and publication rights would be agreed separately.