Start a Project

Global Head of Generative AI, Commercial & Medical

Short description:
  • I scale GenAI copilots (content, insights, QA) with OPDP/HIPAA guardrails—integrated with Veeva/CRM/analytics—to cut rework, speed approvals, and deliver decision-ready readouts.

Who this helps

  • (at a glance)
  • Brand & Commercial Leads — faster claims-grounded content, fewer redlines, clearer ROI.
  • Medical/MLR & Compliance — reviewer packets built automatically with sources and diffs.
  • Analytics/BI — weekly decision one-pagers auto-generated from governed KPIs.
  • Digital/IT/InfoSec — private endpoints, role-based access, no model training on client data.

  • Creative/Production — social-first templates, cut-downs (6/15/30s), captioning, and ISI overlays.

Proof / Mini-KPIs

Time-to-approval ↓

Resubmits ↓

via channel pre-flight

Decision speed ↑

(exec one-pagers auto)

Rework ↓

(acceptance tests)

Auditability ↑

 (source-linked, logged, diffed)

What you get

  • (ready to use)
  • LinkedIn HCP Playbook — audience design (titles/specialties/skills, company filters), un/branded rules, cadence & sequencing, Document/Thought-Leader/Video ad specs.
  • Meta Patient Playbook — Meta Patient Playbook — 18+ gating, sensitive-attribute guardrails, creative DOs/DON’Ts, Stories/Reels/Feed formats, comments policy.
  • Claims & Creative Matrix — copy blocks mapped to evidence; fair-balance & disclaimer placements by format/locale.
  • Pre-flight & Tagging — OPDP/APLB checks, platform-policy screen, UTM & GA4 specs, consent logic, server-side where appropriate.
  • Moderation & PV/AE SOP — response grid, escalation ladder, 24-hour safety handoff, audit logs.
  • Measurement Suite — KPI layer (qualified clicks/conversions), attention & video completion, weekly Decision One-Pager, MMM/MTA-ready exports.
  • Template & Asset Pack — social-first video/image templates, subtitle files, alt text, ISI overlays, localization variants.

The mandate

  • (why this exists)

Your brand, medical, and field teams need faster content, fewer redlines, and tighter readouts—without OPDP/APLB issues or HIPAA headaches. I design and run a compliant GenAI program that plugs into Veeva, CRM, and analytics, so work ships sooner and stands up under review.

Copilot catalog

  • (built for Commercial & Medical)
  • Brand Content Copilot — drafts HCP/DTC copy from approved claims only; annotates qualifiers/timepoints; inserts ISI cues; exports copy decks.
  • MLR Reviewer Assistant — assembles reviewer packets (clean/annotated copy, claims matrix, mockups, references, change log); tracks redlines.
  • QA Copilot (digital) — pre-flight for search/social/banner/email/LP: character counts, ISI adjacency, platform-policy checks, UTM/tagging presence.
  • Medical Evidence Copilot — summarizes label sections and pivotal endpoints; highlights limitations; creates fair-balance snippets.
  • Field/MA Copilot — converts scientific decks into portal modules and compliant follow-ups; logs changes for audit.


Guardrails

  • (baked in, not bolted on)
  • OPDP/APLB — on-label/CFL logic, banned-phrase filters (“cure/safe/best”), short-form rules, major statement + ISI placement patterns.
  • HIPAA/PHI — no marketing tags/retargeting in authenticated contexts; PHI classifiers; BAAs with POC/EHR vendors; aggregate-only reporting.
  • Provenance & audit — retrieval shows exact sources (Label §§, Trial IDs); immutable logs; versioned prompts.
  • Human-in-the-loop — approval gates for publishing; dev/stage/prod separation; least-privilege access.
  • Security — private endpoints, KMS encryption, RBAC; no training of foundation models on your data.

Integrations we wire up

  • Veeva Vault PromoMats (claims library, packet archive), Veeva CRM / Salesforce (approved content, call notes).
  • Jira/ServiceNow (change control), GA4 (consent-aware), Looker/Power BI/Tableau, Snowflake/Databricks (aggregates only).
  • CMP/GTM, CMS, and ad platforms (Google/Meta) with healthcare cert/authorization tracking.
  • Collab in SharePoint/Teams/Slack for assignments and approvals.

Engagements

  • (pick your entry point)
1. Strategy Sprint

Deliverables

  • GenAI opportunity map
  • Risk register
  • Prompt Governance policy
  • Model cards
  • Prioritized backlog

Artifacts

GenAI RACI, Risk/Issues log, Copilot value model.

2. Pilot
  • Stand up two copilots (e.g., Content + QA) with acceptance tests
  • Data contracts
  • Reviewer-packet automation

Artifacts

Claims/Evidence Matrix, Channel pre-flight suites, Packet generator, Decision One-Pager template.

3. Scale & Run
  • Add roles (Medical/Field)
  • Expand to omnichannel
  • Harden SLAs
  • integrate POC signals (RTBC/ePA) into exec readouts

Artifacts

Operating Handbook, Change-control SOP, Evaluation harness & red-team scripts.

KPIs we sign up for

MLR first-pass acceptance • Copy cycle time (brief→approval) • Redlines/asset (and severity) • Packet completeness (0 missing) • Exec time-to-insight (Weekly One-Pager lead time)

What we won’t do

Generate off-label claims or imply superiority without evidence • Touch PHI in marketing analytics or enable patient-level targeting • Write back to Veeva/CRM without a human approver

Book a 30-min Scoping Call
 Download the Service One-Pager

FAQs

Flexible—Azure OpenAI/Vertex/private endpoints with InfoSec sign-off; retrieval runs in your VPC; no base-model training on your data.

Every output shows sources/qualifiers and a diff to the approved claim, plus a read-only packet they can check in seconds.

Weekly Decision One-Pager with the KPIs above; formal evaluation harness (accuracy, grounding, toxicity, privacy).