EIOPA market survey
Where generative AI is taking hold
Use and plans among surveyed insurers
347 insurance undertakings across 25 countries
Reported plans are intentions, and active use can include early experiments.
Show data table
| Category | Share |
|---|---|
| Already using generative AI | 65% |
| Planning to use it | 23% |
Adoption and maturity answer different questions
EIOPA found broad use across respondents, while almost two thirds of the reported applications were still at the proof-of-concept or experimentation stage. One insurer can run several pilots, so the share of insurers using generative AI and the share of applications in production describe different things.
Start with a task that has a clear reviewer
For claims intake or underwriting preparation, select one document type and test against a representative set of completed cases. Measure preparation time, correction effort and the share of files that pass review. Include unusual cases and missing documents so a faster average does not hide failures that require escalation.
Set the decision before funding the pilot
Agree who owns acceptance, which checks remain manual and how the team will roll back. Compare cost per accepted file with your baseline, including review, rework, software and operating costs. Review quality and escalation rates alongside cost before extending the workflow.
Sources and methodology
Genellipse analysed published aggregate figures from EIOPA’s July 2025 market survey, published on 2 February 2026. The charts use the reported percentages and keep the survey population and reported application examples separate. The data describes respondents in participating European markets. It does not establish global adoption, effectiveness or expected returns.
Read the original EIOPA reportChoose a workflow to evaluate
In the first conversation, we’ll identify a workflow owner, available evidence and the criteria for deciding whether a pilot is worth pursuing.
Discuss an insurance AI pilotExplore insurance workflows