Univé builds an AI-ready workforce
OpenAI
Dutch insurer Univé rolled out ChatGPT Enterprise across its whole staff. The twist: they let employees, not just IT, drive how it gets used.
Based on reporting by OpenAI — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Univé isn't the kind of company that shows up in AI headlines. It's a Dutch insurance cooperative, the sort of institution built on decades of careful underwriting and even more careful compliance. So when OpenAI holds it up as a case study in building an 'AI-ready workforce,' the interesting part isn't the technology itself — it's the way Univé chose to introduce it.
Rather than handing ChatGPT Enterprise to a small pilot team or burying it inside an innovation lab, Univé pushed adoption company-wide and paired it with something a lot of enterprise rollouts skip: real governance. Leadership set the tone from the top, laying out where AI could be trusted and where humans still had to sign off, which matters enormously in an industry where a wrong answer can mean a denied claim or a regulatory headache.
But the more telling move was letting employees themselves shape how the tool gets used day to day. Instead of a rigid, top-down list of approved use cases, Univé leaned on staff to surface the workflows worth automating — drafting customer correspondence, summarizing policy documents, that kind of unglamorous but constant work that eats up an insurer's week. That bottom-up input turned into the actual playbook for scaling adoption across teams.
The result, according to OpenAI, is a workforce that treats ChatGPT as a normal part of the job rather than a novelty bolted onto it. And that's really the quiet lesson here: getting large organizations to use AI well has less to do with model capability and more to do with whether the people doing the work were ever asked what they needed it for.
My take — AI-written commentary, not fact-checked reporting
I'll say the obvious thing nobody in enterprise AI wants to say out loud: most rollouts fail not because the model is weak but because some exec mandates it from a slide deck. Univé's approach — governance plus asking the actual employees — is boring, unsexy, and probably why it'll actually stick. European firms get mocked for moving slow on AI, but slow-and-governed beats fast-and-abandoned every time.
Read more about this at: OpenAI