How Moderna made building your own AI part of everyone's job
Moderna didn't just buy ChatGPT Enterprise — it taught thousands of employees to build their own AI assistants, and merged its HR and technology functions to manage a workforce that's part human, part GPT.
After the vaccine boom, a bet on doing more without hiring more
As post-pandemic demand for its COVID vaccine fell, Moderna leaned into efficiency and pipeline acceleration while holding headcount roughly flat. CEO Stéphane Bancel positioned AI not as an IT side-project but as mission-critical — a way for AI-augmented employees to expand output, run more mRNA programs, and move faster without proportional hiring.
The strategy treated AI literacy as a core competency for every eligible employee, not a specialist skill confined to a data-science team.
Pick the tool by who actually uses it
Moderna first built an internal assistant called mChat on OpenAI's API, then ran real user testing comparing mChat, Microsoft Copilot, and ChatGPT Enterprise. It chose ChatGPT Enterprise for the highest user satisfaction and deployed it company-wide.
That ordering matters: the tool decision was driven by which option people actually adopted and liked, not by a top-down procurement preference. Adoption was the selection criterion, not an afterthought.
Non-engineers building their own assistants
The real shift was letting non-technical employees build custom GPTs for their own work — in legal, research, manufacturing, and commercial functions. One example, “Dose ID,” was built to help evaluate optimal vaccine dosing with rationale, source references, and charts. An internal AI Academy trained employees and drove adoption.
The numbers were striking: roughly 750 custom GPTs created in about two months, an average of around 120 ChatGPT conversations per user per week, 40% of weekly active users having built their own GPT, and the legal team reaching essentially 100% adoption. This wasn't a pilot with a few enthusiasts — it was broad, hands-on usage.
Merging HR and technology
Moderna went a step most companies haven't: it merged its HR and Technology functions into one organization under a Chief People and Digital Technology Officer, Tracey Franklin, explicitly to redesign roles around what humans and AI each do best.
It hasn't been frictionless — Moderna has also trimmed some digital-technology roles along the way — but the structural signal is clear: the company treats “who does what, human or AI” as an org-design question, not just a tooling one.
Treating AI proficiency as an operating metric
Moderna is a flagship example of an “AI-native” enterprise: not merely buying tools, but restructuring around a human-AI division of labor and democratizing who gets to build AI agents. The trend it points to — employee-built assistants and workforce AI proficiency tracked like any other operating metric — is exactly where Fautons argues adoption has to go.
The headline isn't the 750 GPTs. It's that fluency with AI became part of everyone's job description, and the organization was rebuilt to make that real.
The shift
- AI as an IT pilot, owned by a few
- One-size tool chosen top-down
- AI proficiency left to chance
- HR and technology run as separate functions
- Output tied tightly to headcount
- AI mission-critical, owned company-wide
- Tool chosen by real user satisfaction
- Employees building ~750 GPTs in two months
- HR + tech merged around human-AI work
- Proficiency treated as an operating metric
Moderna's real move wasn't buying AI — it was making fluency with it part of everyone's job, and redesigning the organization around that.