Shift Technology introduced the ARISE framework, a transformative standard designed to clarify the deployment of agentic AI within the insurance ecosystem. This groundbreaking methodology establishes a clear roadmap for how insurers can safely measure, trust, and scale AI autonomy throughout the claims lifecycle.
Eric Sibony, Shift Co-Founder & Chief Scientific Officer, introduces the ARISE framework
Agentic AI brings a massive transformation opportunity for insurance, but with so much buzz around the technology, carriers are often left wondering what it practically means for their operations. That is why we introduced ARISE: a definitive standard defining five distinct levels of AI agent autonomy built specifically for insurance use cases.
The framework breaks down as follows:
-
Level 1: Answers. The agent operates within a conversational interface, answering specific inquiries from claim handlers.
-
Level 2: Recommends. Moving beyond passive answers, the agent proactively analyses data to highlight insights and recommend the best next steps for complex claims without being prompted.
-
Level 3: Initiates. While keeping a human in the loop, the agent prepares all necessary actions—such as drafting formatted emails—leaving the handler with just a single click to validate and send.
-
Level 4: Solves. At this stage, the agent processes and solves claims in full autonomy. Reaching this requires strict reliability, achieving over 99% accuracy by learning from the systematic oversight of Level 3.
-
Level 5: Exceeds. The ultimate state of insurance AI. Operating autonomously, the agent mimics top-tier claims handlers by knowing when to safely deviate from standard operating procedures to optimise the final claim outcome and maximise customer satisfaction.
Ultimately, ARISE isn't just about driving operational efficiency or accelerating claims processing. It is about equipping insurers with a reliable, standardised vocabulary to deploy AI that makes better decisions, manages indemnity risk, and delivers superior outcomes.
