SingaporeEnergy and utilities2025 – ongoingPuma Energy
Predictive stockout alerts and generative AI audit preparation
Puma Energy cut fuel stockouts by more than 20% and shortened audit preparation from two weeks to one hour
Problem
Puma Energy runs fuel stations in more than 20 countries, many in underserved communities. Its data sat in separate regions, and teams worked from offline exports. Safety and audit preparation took two weeks.
Approach
Puma Energy moved its data to the Databricks platform with shared governance across regions. On top of it, it built predictive alerts for fuel stockouts. It also used generative AI for audit and safety preparation, invoice and document processing, and a retrieval-based analysis of customer feedback.
Results
Takeaway
Puma Energy's stockout and audit-preparation results came on top of a data platform migration. The company did not say how much of each result came from the models and how much from the cleaner data.
Categories
- Size
- Large enterprise
- Industry
- Energy and utilities
- Business function
- Production and qualityLogistics and supply chain
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and MLGenerative AI
- Adoption
- Off-the-shelf tool or platform
- Evidence
- Vendor case study
Sources
Source summaries
Databricks describes Puma Energy's move to a unified data platform and the machine learning and generative AI uses built on it. It reports a stockout reduction of more than 20% from predictive alerts and audit preparation cut from two weeks to one hour. It also reports a 98 to 99% drop in data latency and 15% data cost savings, which come from the platform migration.
Full details are in the original sources.
