Insurance data conference wraps up FASI’s 50th anniversary

Bayes Business School hosted the Insurance Data Science Conference 2025 (IDSC) last month.

Co-organiser Andreas Tsanakas, Professor of Risk Management at Bayes and co-founder of the event, said:  “The conference brought together leading experts from academia and industry worldwide to share ideas highlighting how innovation is happening globally in this fast-changing field. We are proud to provide a platform for academics to present their research and discuss with industry practitioners how theoretical breakthroughs in data science can be effectively translated into practical solutions for insurers.”

“The IDSC is the premier international forum for advancing insurance data science and this event really underscored the transformative power of data science and AI in the insurance sector.”

A core theme of the conference was the critical collaboration between academics and insurance practitioners, working together to explore research frontiers and industry applications.

Thematic highlights from the presentations included:

• Advanced regression methods – evidencing a wide range of approaches and innovative problem-focused architecture including tree-based, neural, graphical and Bayesian models
• Pricing and portfolio management – ranging from sophisticated uses of market data to market analytics
• The deployment of language models and AI in applications ranging from claims processing to data granularity improvement and risk identification
• Sustainability and climate risks – including the design of insurance indices and the advanced modelling of climate effects
• Fairness and 'explainability' – from measuring discriminatory pricing effects to designing interpretability into model architectures.

All talks and abstracts are available in the conference programme.

Inspiring Keynote Speakers

Three distinguished keynote speakers shared their expertise and vision for the future of insurance data science. Magdalena Ramada from WTW opened the conference by discussing the current impact of emerging technologies on the insurance value chain. James David Long, CTO of Palomar, explored whether data engineering is the new data science. Johanna Ziegel from ETH Zurich delivered insights on conformal calibration guarantees for reliable predictions, offering technical tools for assessing uncertainty and building trust in predictive models.

Their presentations provided a high-level view of strategic directions and critical considerations within the field.