Field notes
What we think is breaking, before the industry admits it.
A point of view on where pharma marketing is exposed: AI search, the MLR bottleneck, message fidelity across channels, and the content supply chain. No recaps of last year. The shift that is happening now, and what to do about it.
01/Reports
White papers
Long-form, citable writing on the problems Juncture was built to close. Read the summary on the page; the full report is one short form away.
- For medical, regulatory, and commercial leaders
The Strategy Execution Gap
Why pharma loses the approved message after approval: four lifecycle friction points, the four controls that solve parts but not the connection, and how to govern message execution on one approved source of truth.
Read the report - For medical affairs and content operations
Reference Quality in Pharma Claims Libraries
A five-check framework for the citations nobody verifies: reference present, section substantiates, numbers match, label version current, evidence hierarchy respected, with a library-wide health rollup.
Read the report - For commercial and medical affairs
The Off-Label Drift Report
How AI answer engines pull approved messages off-label, where the drift starts, and how to measure it across ChatGPT, Gemini, Perplexity, Google AI Overviews, and Claude.
Read the report - For MLR, regulatory, and brand teams
The Pre-MLR Pre-Check Playbook
A working method for catching message, fair balance, and ISI problems before an asset reaches MLR, so reviewers see a cleaner asset and a defensible record.
Read the report - For brand, medical, and analytics teams
The Six Core KPIs for Pharma Share of Answer
How to define and compute the six Core KPIs of pharma Share of Answer, with the denominator problem, per-engine sampling, and a worked example across the answer engines.
Read the report - For content ops and brand teams
The Content Intelligence Playbook
How to turn an approved-content library and reuse scoring into measurable AI visibility, and close the loop with AI Pickup and Answer Monitor.
Read the report
02/Field notes
Field notes
- pharma claims library reference quality9 min read
Reference quality: the silent risk in your pharma claims library
Pharma claims library reference quality: five checks for missing, stale, or non-substantiating references, and why label changes make currency a moving target.
Read - semantic content reuse pharma8 min read
Why Exact-Match Reuse Detection Misses Your Own Approved Content
Exact-match reuse detection scores approved pharma content as new the moment a word changes. Semantic content reuse matching resolves edits back to the module.
Read - HCP question set AI monitoring8 min read
How to Build an HCP Question Set: Voices, Moments, and Why Keyword Lists Fail
Keyword lists fail as AI probes. Build an HCP question set for AI monitoring from voices crossed with moments: six probe families, one guardrail, governance.
Read - AI answer readiness pharma9 min read
AI Answer Readiness: Can an Answer Engine Actually Read Your Label?
AI answer readiness pharma: an auditable, five-category check of whether an answer engine can fetch, parse, and trust your label before it writes.
Read - answer operating model8 min read
The Answer Operating Model: Who Owns What AI Says About Your Drug
When an AI assistant characterizes your therapy to an HCP, that answer is promotion that nobody owns. The answer operating model fixes ownership on one approved label.
Read - clinical AI engines8 min read
The Clinical AI Engines Where Prescribing Decisions Actually Happen
HCPs use clinical AI engines like OpenEvidence at the point of care. Why pharma must monitor clinical AI engines, not just consumer chatbots, against the label.
Read - pharma message integrity AI8 min read
Pharma Message Integrity in the Age of AI: Closing the Gap Between Approved Label and Model Answer
Approved pharma claims and live AI answers drift apart. See why MLR tools and AI-visibility tools each see half, and how the join closes the message-integrity gap.
Read - pharma share of answer KPIs8 min read
The Six Pharma Share of Answer KPIs for AI Engines
A plain-language guide to the six pharma share of answer KPIs (Share of Answer, Ecosystem SoA, Precision, Risk, Claim Uptake, Top References) for measuring AI engines.
Read - pharma claims library7 min read
The Pharma Claims Library: Your Approved-Content System of Record in the AI Era
A pharma claims library is the approved-content system of record that anchors AI visibility. Why scattered content fails answer engines, and how reuse scoring and semantic where-used fix it.
Read - AI content pickup pharma8 min read
AI Pickup: how much of your approved content the machine actually echoes back
AI content pickup in pharma measures how much MLR-approved content the answer engines echo to HCPs per engine. Invisible approved content is wasted MLR effort. How to measure echo and exposure.
Read - pre-MLR content check7 min read
Pre-MLR Content Check: What a Cleared, Review, or Blocked Verdict Actually Means
A pre-MLR content check returns Cleared, Review, or Blocked by grounding every claim in the label of record, not a brand book. Here is how the verdict works.
Read - monitor ai answers9 min read
How to Monitor AI Answers About Your Brand (Without Sending the Models Anything Confidential)
How to monitor AI answers about your brand across ChatGPT, Gemini, Perplexity and AI Overviews: what to probe, how to score, and how to track answer share.
Read - GEO for medical devices11 min read
GEO for Regulated Healthcare: What Pharma, Medtech, and Medical Devices Share When AI Answers the Question
GEO for medical devices, medtech, and pharma is one problem: a regulated claim, paraphrased by a model, can drift off the cleared indication or IFU.
Read - FDA AI pharma enforcement 202511 min read
FDA Went From 5 Enforcement Letters to Over 100: What the 2025 Crackdown Means for AI-Assisted Pharma Content
FDA AI pharma enforcement 2025 jumped from 5 OPDP letters in 2 years to 100+ in a day. The trigger was misleading risk claims, not the AI that wrote them.
Read - AI in MLR review pharma12 min read
Can You Actually Use AI in MLR Review? A Decision-Grade Read of the Rules
AI in MLR review pharma is allowed in exactly one shape: an augmenting pre-check that flags issues for human reviewers, never a black-box approver.
Read - HCP AI search behavior9 min read
HCPs already switched to AI search. Here is the data pharma keeps ignoring.
Physicians and patients moved from search and reps to AI assistants for drug answers. The cited data on HCP AI search behavior, why it strands your approved message, and what to instrument.
Read - AI pre-check before MLR pharma8 min read
What an AI pre-check actually checks before MLR, and what it must never do
A concrete walk through a pre-MLR pre-check in pharma: how it extracts claims, runs deterministic fair balance, ISI, on-label, and reference rules, cites the label, and where its hard guardrails sit.
Read - claim drift RAG9 min read
How retrieval and summarization quietly rewrite your approved claim
Claim drift is structural. An answer engine retrieves mixed sources then summarizes, dropping a qualifier so an on-label claim reads off-label. Here is the mechanism, step by step.
Read - approved content AI visibility pharma8 min read
Your approved content is invisible to the machine. Here is how to start closing the gap.
Pharma brands sit on libraries of clause-cited, MLR-approved modules the AI never cites, so third-party sources win the answer. How to make approved content retrievable, measure what the machine quotes, and feed the gaps back.
Read - GEO vs AEO vs LLMO pharma11 min read
GEO vs AEO vs LLMO for Pharma: A Plain-English Field Guide to the New Search Vocabulary
GEO vs AEO vs LLMO in pharma is not three rival acronyms. It is one layered job that only works on a compliance backbone the generic playbooks skip.
Read - ChatGPT drug information accuracy7 min read
A 1-in-3 Chance ChatGPT Gets Your Drug Wrong: The Accuracy Numbers Pharma Leaders Need
ChatGPT drug information accuracy is worse than most brand teams assume. The published record: roughly 74 percent of drug answers wrong, citations routinely fabricated.
Read - AI Overviews health searches pharma7 min read
AI Overviews Now Answer Roughly Half of Health Searches Before the Patient Reaches Your Site
AI Overviews health searches in pharma now near 51%. The answer is assembled before the click, making AI visibility a brand KPI and a compliance surface.
Read - modular content pharma MLR7 min read
Modular Content Plus a Pre-Check: The Two-Part Fix for the 50-60 Day MLR Cycle
Modular content pharma MLR works best with a pre-check: governed modules and clean new blocks compound, so cycle-time savings multiply, not add.
Read - agentic AI pharma marketing8 min read
From Omnichannel Strategy to Orchestration: What Agentic AI Changes for Pharma Commercial Teams in 2026
Agentic AI in pharma marketing is moving from analysis to action in 2026. Why orchestration without a message-fidelity loop is just faster risk.
Read - share of answer9 min read
Share of Answer is the metric that replaces share of voice. Most pharma teams cannot measure it yet
Share of Answer is the percentage of AI answers that mention your brand, accurately. Here is how to define it, instrument it, and why it replaces share of voice in pharma.
Read - generative engine optimization pharma9 min read
GEO is the new SEO, and pharma is the least prepared industry for it
Generative engine optimization in pharma is not a 2027 problem. HCPs and patients are asking AI, visibility now means Share of Answer, and regulated brands are exposed.
Read - AI off-label pharma9 min read
Your approved message is already off-label in the machine answer
AI off-label pharma exposure is real and invisible. When HCPs ask ChatGPT or Perplexity about your drug, the model paraphrases your label and reads off-label. Here is the mechanism.
Read - MLR review bottleneck9 min read
MLR was always the bottleneck. AI did not fix it. It made it worse.
GenAI lets pharma make 5 to 10x more content, but the MLR review bottleneck cannot scale to match. The fix is a pre-check before review, not faster review.
Read - omnichannel pharma9 min read
Omnichannel measured reach. It never checked whether the machine got the message right
Omnichannel pharma optimized impressions and share of voice but never message fidelity. As HCPs move to AI answers, reach is vanity. Fidelity is the goal.
Read - pharma content supply chain9 min read
The pharma content supply chain was built for a world that is ending
The pharma content supply chain was built to feed human channels. In an AI-answer world, claim fidelity and content reuse become the metrics that matter.
Read
See it on your brand
Stop reading about the gap. Watch it close.
Bring an asset and a brand. We will pre-check the asset against the label and show how the machine answers about the brand today.