CyberArk's support-case automation

An AWS technical post co-written with a CyberArk engineer describes an AI-assisted pipeline that parses varied customer logs using Amazon Bedrock, then makes them queryable in Apache Iceberg. The authors report up to a 95% reduction in the time from support-case assignment to resolution: simple cases reportedly moved from 4–6 hours to 15–30 minutes, and complex cases from as long as 15 days to 2–4 hours. The architecture also replaced manual onboarding of new log formats. These results are customer/vendor-reported, not independently verified; the reported "up to" figure must not be read as an average.

This is a concrete deployment of AI inside a support workflow, not just a model purchase. The implementation combined data engineering, privacy masking, dynamic parsing, and a serverless processing path.

Related

• pythian-gemini-enterprise-database-operations — another case of AI reducing incident-resolution time.
• regal-rexnord-copilot-agents — support knowledge agents serving employees and customers.