LinkedIn link GitHub link

Mirai at Insurance Data Science Conference 2025

Back to overview

Grateful for the opportunity to contribute at the Insurance Data Science Conference 2025.

Group Photo, credit IDS
Source IDS 2025.

The Insurance Data Science Conference 2025 (IDSC 2025) took place in London on June 19-20, bringing together researchers and practitioners to discuss the latest advancements in data science and AI within the insurance sector.

The conference facilitated sharing insights, networking, and exploring innovative insurance solutions through inspiring keynotes, an Academia-Industry panel, and diverse technical sessions.

Leading Innovations in Data Science and AI

Presenters shared innovative research and hands-on applications driving change across the insurance landscape:

  • Practical use of natural language processing, with applications ranging from claims processing to risk assessment.
  • Advanced regression methods, including Bayesian and graphical models, neural networks and auto-calibration.
  • AI/ML applications in a variety of domains, such as fraud detection, loss prediction, reserving, pricing and claims modeling.
  • Fairness and explainability, from measuring discriminatory pricing to improving explainability of opaque ML models.
  • Climate risk modeling, showcasing various approaches to advanced modeling of climate effects.

Agentic AI Applications in Insurance

The conference also featured a session on Agentic AI Applications in Insurance, where Mirai Solutions presented our insights on the use of agentic AI in the insurance industry. We discussed not just the potential benefits of agentic AI, but also the challenges and risks associated with its implementation. We highlighted the importance of adopting instrumental design patterns, like Evaluation-Driven Development, check and balance mechanisms, and a robust data strategy. Finally, we introduced a strategic roadmap to guide insurers in building agentic AI applications, from assessing use cases to enabling continuous improvement.