Univé, one of the Netherlands' largest insurance cooperatives, has launched a structured, organization-wide program to train its workforce on AI tools, according to a case study published by OpenAI. The initiative is framed around a common but under-discussed problem: giving employees access to AI tools does not automatically translate into productive use.
According to the OpenAI writeup, Univé's approach centers on role-specific training rather than blanket rollouts of generic AI literacy content. The company reportedly worked to identify where AI could meaningfully change day-to-day work across departments — including customer-facing and back-office functions — and built training paths around those specific use cases, rather than a one-size-fits-all onboarding session. Details on exact headcount trained, timeline, or measured productivity outcomes were not fully specified in the source material and should be treated as unconfirmed pending further reporting.
This distinction — tool access versus tool fluency — is one that has surfaced repeatedly across enterprise AI rollouts over the past two years. Numerous surveys and case studies from vendors and consultancies have pointed to a persistent gap: companies purchase AI licenses at scale, but usage rates and reported productivity gains lag far behind expectations. Univé's program appears to be a direct response to that pattern, treating structured training as a prerequisite for return on AI investment rather than an optional add-on.
For large enterprises like Univé, with thousands of employees across multiple departments, building a dedicated internal training function makes sense — they have the scale to justify a standing program and the internal L&D resources to run it. That's a meaningfully different situation from a company with a few dozen or a couple hundred employees, where there's no room for a dedicated AI training department and no time for months-long rollout programs.
What transfers from the Univé case to smaller organizations isn't the scale of the program but the underlying diagnosis: AI tool access without workflow-specific training tends to underperform. A support team with an AI assistant that isn't tuned to your ticket categories, tone, and escalation rules will use it inconsistently or not at all. A sales team with AI drafting tools that don't reflect your actual proposal structure will revert to manual work. The fix isn't necessarily a formal training curriculum — for smaller teams it's often a handful of concrete workflow integrations, reviewed and refined over a few weeks, rather than a company-wide learning initiative.
Univé's case adds to a growing body of evidence that the constraint on AI value in the workplace is increasingly organizational and procedural, not technical. The tools exist and are widely licensed; the gap is in deliberate integration into specific, repeatable tasks. Companies evaluating their own AI investments — at any size — may find more value in auditing existing tool usage than in acquiring additional capabilities.