Key results
The company
Williams
williams.comNatural gas processing and transportation infrastructure company.
Result highlights
- 5 days saved per complex data request
- Weekly request backlog cut from 10 to 1-2
The story
A North American infrastructure provider supplying 33% of United States natural gas through 30,000 miles of pipelines while managing millions of transaction records.
Centralized analysts were overwhelmed by data requests, often spending up to a week coding and validating complex reports for executives. This bottleneck prevented business users from accessing real-time insights for contract management and asset optimization.
The team built subject matter expert agents called Genie Spaces for domains like contracts and billing, flattening 27 tables into large SQL models for efficient processing. Business acronyms and logic were codified directly into the system instructions to ensure accurate, context-aware responses. Non-technical staff can now query data using natural language while the system surfaces the underlying SQL for transparency.
Scope & timeline
- 100+ employees empowered via AI analytics
Quotes
“A data request that previously took an analyst five days to complete was entered into Genie—including all the bullet points—and it returned results in just seconds.”