You are Chiara Beltrami, Chief Strategy Officer of Assicura Verde, a Milan insurer with EUR 90 million and three years to transform it. A digital rival is taking motor business, the supervisor has just written to the sector about AI pricing, and your data team wants a new product in market within eight weeks. Before a euro is spent, you must decide what part the rulebook plays in your strategy.
The situation
It is Monday, 2 February 2026. Assicura Verde, founded in 1961, insures cars, homes, health and small businesses for 2.4 million customers. It wrote EUR 3.1 billion of premium last year, with a 96% combined ratio and a 178% Solvency II ratio. Its defining feature is the network: 1,850 tied agencies sell 78% of premium, about EUR 2.4 billion.
Lumo, a digital-native insurer, has taken 9% of new motor policies in your northern markets in eighteen months. It quotes a driver in under 3 minutes; your website takes 14. The board has given you EUR 90 million over three years, none of it committed yet, and one instruction from the chair: modernise without tripping the regulator and without breaking the agency network.
This same week IVASS, the Italian supervisor, warned the sector that AI underwriting must meet the EU AI Act's high-risk obligations (documentation, human oversight, bias testing) before it goes live. Davide Orsini, head of data science, wants a usage-based motor product launched for the spring renewal season, which closes in eight weeks, and credits it with four points of combined-ratio improvement. Franco Tessaro and a second underwriter, with twenty-five years of claims each, think the model was fitted to a benign period and will misprice young drivers. The general counsel estimates that a launch IVASS judges to have gone live without those obligations would cost EUR 12 to 18 million in remediation, plus a Solvency II capital add-on.