Cognitive claims handling with AI
summary
Company builds a narrow AI claims engine that combines computer vision and natural language processing to extract about 200 data fields from a claim (vs 7–14 with legacy tools). Platform trains on a massive corpus (started with 30 million documents and ~980 million digitized over time), offers pre‑trained models that need only a few hundred customer examples to fine‑tune, runs fast retrain cycles, and is live with insurers delivering automated decisions, recommendations and measurable speed and transparency gains.
AI-written summary · about
chapters (19)
- 0:00Intro
- 0:58The human brain
- 4:42Where are we
- 6:15What are the arguments
- 6:54Everything is different today
- 7:26Focus and use case
- 10:40Why focus
- 11:30Why is it exciting for us
- 14:42We have a clear mission
- 15:17From process to datadriven
- 15:27Give back time
- 19:10Algorithms
- 19:55Uncomfort zone
- 20:38Customers
- 20:49What would be different
- 20:59More human
- 21:18We are convinced
- 21:48War
- 22:04Questions
related talks
22mHarnessing AI for Underwriting in Insurance
13mTransforming Insurance Through AI
50mWhy use AI in insurance underwriting and claims? (PRT271)
16mAI-Driven Claims Adjudication: Healthcare Efficiency & Fraud Prevention
09mAI-Driven Insurance Operations
16mGenerative AI for Claims and Fraud
41mAgentic AI and the Human Role Illustrated by Claims Handling
35mGenerative AI for Fraud Detection & Claims Analysis