Pfizer
genAI marketing-content production platform under regulatory control
In February 2024, Pfizer launched Charlie, a generative AI platform built with Publicis on the Marcel base, which generates emails, digital media, and sales presentations under a red/yellow/green review system, deployed from 600 beta users to thousands of people across its brands.
Objective
Speed up the production and review of Pfizer's marketing content (emails, media, sales presentations) by giving brand teams and agencies a single tool that generates, fact-checks, and pre-sorts what medical review needs to examine closely.
The deployment
Charlie, named after co-founder Charles Pfizer, is a generative AI platform launched in February 2024 for the marketing-content production chain. Built by Publicis Groupe on the base of its Marcel platform, it generates digital media, emails, and presentations that sales teams use with physicians, as well as drafts of medical articles. The content is generated via a customized version of GPT trained on already-approved content classified by therapeutic area and by product, and the answers are cross-checked against those sources to limit hallucinations. A red/yellow/green risk system pre-sorts the assets based on the attention that medical review must give them. At launch, about 600 beta users from the central marketing team were using it, with an extension to thousands of people across the brands and to the Publicis and IPG partner agencies.
Results Proof C
The deployment is documented by the press (Digiday) and by a second consistent publication, with a quantified adoption scale (600 beta users, thousands across the brands). No financial result or independent study on the impact, hence C rather than A or B.
How it works
Documented architectureThe stack in detail
- llm GPT customise (OpenAI) Customized version of GPT trained on Pfizer's already-approved content, classified by therapeutic area and by product, with answers cross-checked against those sources.
- plateforme Publicis Marcel Publicis Groupe platform on the base of which Charlie was built.
- integrateur Publicis Groupe Builder of Charlie for Pfizer; the Publicis and IPG agencies also use the platform.
- outil Systeme de pre-tri du risque rouge/jaune/vert Custom Charlie component that classifies each generated asset by the attention that medical and regulatory review should give it.
How it runs, concretely
For ops teams-
1Framing the content need marketing
The brand team describes the asset to produce (email, media, presentation) for a given therapeutic area.
-
2Generation AI
Charlie generates the content from the approved corpus and cross-checks the claims against the sources.
-
3Risk pre-sorting AI
The red/yellow/green system flags the assets that require the most attention in review.
-
4Medical and regulatory review medical and regulatory team
Review examines the higher-risk assets first, before approval.
-
5Distribution marketing
The approved content is distributed through the marketing channels and by sales reps to physicians.
The corpus of already-approved content, classified by therapeutic area and by product, on which the model relies. Without this up-to-date corpus and without human review, the risk of a non-compliant claim breaks the setup.
How your customers perceive this type of use
Sourced studiesUn ecart net separe les annonceurs des consommateurs : 77% des annonceurs voient l'IA positivement contre 38% des consommateurs (Yahoo/Publicis, 2024). Les mesures implicites confirment le rejet declare : en EEG, les pubs generees par IA produisent une activation memorielle plus faible que les pubs traditionnelles et sont decrites comme agacantes, ennuyeuses et confuses (NIQ, 2024). La disclosure a un effet ambivalent : elle augmente fortement la confiance quand elle est remarquee (Yahoo/Publicis), mais 27% des jeunes consommateurs disent faire moins confiance a une entreprise dont la pub est creee par IA (IAB, 2024).
Acceptance conditions
- Une disclosure visible : quand la mention IA est remarquee, la confiance globale envers l'entreprise augmente de 96% (Yahoo/Publicis 2024)
- Une qualite visuelle suffisante : les visuels IA de basse qualite augmentent l'effort cognitif et distraient du message (NIQ 2024)
Red lines
- Le contenu IA non declare puis identifie : 72% des consommateurs disent que l'IA rend l'authenticite difficile a etablir (Yahoo/Publicis 2024) et les marques utilisant des pubs IA sont plus souvent jugees inauthentiques ou non ethiques par les consommateurs que par les dirigeants (IAB 2024)
- Les mannequins et personnes generes par IA : 46% des consommateurs n'en veulent pas dans la publicite, l'inquietude premiere etant les standards de beaute irrealistes (Attest 2025)
Sources: Yahoo / Publicis Media (terrain Ebco) 2024 · IAB (avec Attest) 2024 · NIQ (NielsenIQ) 2024 · Attest 2025
How to replicate
Inference, not sourcedData prerequisites
- corpus of already-approved, classified content
- product and therapeutic-area reference
- promotional-review rules
Org prerequisites
- medical and regulatory review team
- marketing-content governance
- AI integration partner
Possible stack
- LLM (GPT or equivalent) with RAG on the approved corpus
- content orchestration platform
- review workflow
The plan, step by step
- Step 1Build the corpus of already-approved content, clean and classified by therapeutic area and by product, with the promotional-review rules by market.Deliverable: RAG-usable corpus + rules reference
- Step 2Connect an LLM to the corpus with systematic cross-checking of claims against the approved sources, on a first asset type (email).Deliverable: Generator in beta on one content format
- Step 3Build the risk pre-sorting (red/yellow/green) and the workflow that prioritizes medical and regulatory review.Deliverable: Tooled, operational review circuit
- Step 4Open a beta to a core of marketing users and measure production speed, correction rate in review, and adoption.Deliverable: Quantified pilot assessment
- Step 5Extend to the other brands and the partner agencies, with training and content governance.Deliverable: Platform generalized across the marketing organization
First step: Build a corpus of approved content, clean and classified, usable in RAG by an LLM under human review.
The same pattern, proven in other industries
Adobe
in-house genAI production of marketing content at scale
Atlassian
internal genAI agents for account targeting and marketing content production
BMW
advertising and experiential creation with generative AI
Sources
- S1 With 'Charlie,' Pfizer is building a new generative AI platform for pharma marketing Established press archive pending
- S2 Pfizer partners with Publicis; launches AI platform Charlie Secondary archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.