back_office_ops · finance · workflow
Travelers Insurance classifies service request emails with Amazon Bedrock and Anthropic Claude, achieving 91% accuracy
Travelers receives millions of emails a year requiring classification of complex, sometimes ambiguous service requests into categories such as address changes, coverage adjustments, and payroll updates — a volume of manual processing that warranted automation.
How it works
Common implementation structure
How this type of workflow is generally built, generalized across documented cases — not tied to any one vendor's stack. Click any stage to read what happens there. Specific products that implement these stages appear in “Tools commonly seen” below.
Stage 1 · Email ingestion
The raw email is ingested into the pipeline and the body text is extracted from the email text files.
Tools used
Amazon BedrockAmazon TextractAnthropic's Claude
Outcome
Through prompt engineering with Anthropic's Claude models on Amazon Bedrock, the team achieved 91% classification accuracy — up from 68% without prompt engineering — with the system positioned to save tens of thousands of hours of manual processing.
Results
Time savedtens of thousands of hours
Volume91%
Cost replaced90%
Grounding & classification
Source type: technical build writeup
25 fields verified against source quotes.
data extractiondocument classificationocremailform submissionmetric backednamed customertools describedvendor confirmedworkflow describedinsuranceaccuracy improvementautomation ratetime savedtechnical build writeupback office opscustomer supportdocument to recordextract classify route