After a year of implementing a cutting-edge generative artificial intelligence (genAI) tool to streamline claims document processing, Sedgwick, a global insurance claims management firm, boasts of the technology’s near-perfect accuracy.
Operating in 80 countries, Sedgwick handles around 1.7 million pages of digital claims-related documents daily. These documents undergo meticulous scrutiny by examiners to determine their validity and appropriate handling.
In April 2023, Sedgwick introduced Sidekick, a genAI tool designed for document summarization, data classification, and claims analysis. Powered by OpenAI’s GPT-4, Sidekick offers the company a vast array of language models tailored for various purposes.
Last December, Computerworld interviewed Leah Cooper, Sedgwick’s global chief digital officer, discussing the challenges and benefits of the genAI rollout. By that time, Sidekick had processed 14,000 documents and impressed with its accurate summaries. Fast forward five months, Sidekick has processed over 50,000 documents, achieving a remarkable 98%+ accuracy rate in summarizations, as reported by over 1,000 examiners.
Leah Cooper, Sedgwick’s global chief digital officer
Sedgwick
Computerworld caught up with Cooper again to delve into the insights gained from utilizing genAI in reducing workloads and enhancing document processing efficiencies.
Cooper shared insights on the types of documents evaluated by Sidekick, mentioning a focus on medical-related insurance documents initially, but later expanding to various document types and photos for validation. The average document length during testing was six to seven pages, with some extending up to 25 to 30 pages.
2024-05-16 09:51:03
Article from www.computerworld.com