ASTIN Webinar: AI in Practice — Building Real Actuarial Tools from Prompts to Production

AI in Practice: Building real actuarial tools, from prompts to production

Part 2: In this second session, we get hands-on. We will present one or two concrete AI tools built for non-life insurance use cases, and walk through how they were actually made: the prompts, the iterations, the trade-offs, and the skills required. We will also address the questions that matter in a regulated environment: security risks, data sharing, and how to implement verification steps. Attendees will leave with a realistic view of what it takes to go from an idea to a working actuarial tool, and of the vigilance points along the way.

Part 1: We explored the fundamentals of AI, key model families and milestones, and how these technologies have evolved in the insurance sector.

Geoffrey Nichil is Co-Founder and CEO of Dotika, a Luxembourg-based AI strategy and delivery firm. He holds a PhD in probability theory and an Executive Master's degree from Ecole Polytechnique. Before Dotika, he spent 10 years leading the data & AI strategy of a major Luxembourg insurer, covering data governance, IT modernisation and AI solutions in production.

Alexandre Hotton is Co-Founder and CAIO of Dotika. He holds a Master's degree in computer science and works mainly on deep learning, open-source LLMs, MLOps and GPU infrastructure. Previously Lead Data Scientist at Foyer Assurances, where he brought generative AI solutions into production at scale, he co-founded Dotika in 2025 and leads its technology and product development.