Mira: Hybrid Logic-AI for Claims

Beschreibung

Claims in regulated sectors, such as insurance, remain inconsistent and costly. Manual reviews, static digital forms and opaque systems create inefficiencies, errors and compliance risks, while regulators increasingly demand transparency and speed. This presents an opportunity for a solution that combines automation with compliance-grade auditability.

Mira is an award-winning, logic-first platform that converts complex policy wording into machine-executable logic. It enables the formalisation of legal text into computable contracts to automatically determine claim eligibility and payout amounts. Every outcome is explainable with reference to the originating clause, making Mira unique among existing claims technologies.

The innovative aim is to now fuse this deterministic foundation with targeted AI. New modules will extract rules from natural language, generate atomic information units and guide user interactions, reducing contract configuration time. Crucially, AI will act as an assistant, while the deterministic engine continues to execute all final decisions. This hybrid approach combines compliance-grade determinism with AI-enabled efficiency, overcoming the rigidity of legacy systems and the opacity of pure AI solutions.

Existing solutions either provide partial auditability or stop short of final determinations. No provider owns the “logic layer” that delivers fully-auditable, automated outcomes. Mira will by defining a new category - 'Claim Determination as a Service' (CDaaS).

The project unites Yarowa AG, ZHAW and Mira, as a carved-out entity. Yarowa provides infrastructure and access to over 100 insurance customers. ZHAW contributes internationally recognised expertise in computational logic and AI. Mira - under future CEO, Reuben John - will drive IP development and commercialisation.

The project ensures long-term national benefit, creating jobs in the private and public sectors, while strengthening Switzerland’s role as a leader in explainable automation.