Genmab's 'AI Everywhere' Push Now Saves Staff 3.5 Hours a Week

Genmab isn’t dabbling in AI — the Danish biotech has gone all-in, scaling ChatGPT Enterprise from a 1,000-employee pilot to a company-wide rollout of more than 2,000 licenses in under a year. The initiative, which they’ve branded ‘AI Everywhere,’ is generating hard productivity numbers that would make any CFO lean in: employees report saving an average of 3.5 hours per week, with power users clocking roughly 120 ChatGPT conversations each over the same period. Tahi Ahmadi, Genmab’s CMO and head of experimental medicines, framed the effort bluntly: ‘Genmab’s ambition is to integrate AI into everything we do,’ targeting gains in scientific quality, decision-making speed, and operational efficiency as the company grapples with the complexity of having recently tripled in size.

What makes the rollout stick isn’t just executive mandate but a surprisingly low-friction onboarding tactic. Genmab simply told its people to paste their job description into ChatGPT and ask, ‘What can you do to help me?’ It’s a clever psychological nudge that sidesteps the blank-page paralysis most users feel. That bottom-up experimentation has since spawned a library of over 100 custom GPTs handling everything from drafting clinical documents and summarizing dense scientific literature to running advanced analytics. The medical affairs team, in particular, has found a niche use case in GPT-4o’s new vision capabilities — the model can now interpret complex, image-heavy documents to fact-check medical publications, a task that used to devour time and money.

Perhaps the most ambitious build is the AI Translator GPT, which effectively clones a human translation team inside the model. It assigns roles like translator, editor, and a specialist in biological jargon, even layering in a reverse-translation step to catch what gets lost and trigger another editing cycle. For English-to-Chinese work, it sometimes routes through Japanese as a staging language to boost accuracy. The result? Jobs that once took weeks now wrap in hours or minutes — and the cost savings are significant, though Genmab hasn’t put a dollar figure on them yet. Clinical study reports, patient safety narratives, and IND applications are also getting the GPT-4o treatment, with the model trained to iterate on its own drafts based on clinical feedback, mimicking a human problem-solver rather than spitting out a one-shot output.

Genmab made a deliberate choice to partner directly with OpenAI rather than going through a cloud provider — a move that gives them access to the research and deployment teams who can help push the limits of what the models can do in a regulated industry. The company stresses that it vetted OpenAI’s security and privacy controls before launch and runs periodic reviews to keep things compliant. But the real signal here isn’t the technology. It’s that a biotech with serious skin in the game is measuring AI impact in hours saved per employee and scaling it enterprise-wide. The open question is whether those 3.5 hours are getting reinvested into deeper science or simply making an overstretched workforce whole again.