Session
Unifying Upstream Intelligence with Generative AI
Overview
| Experience | In Person |
|---|---|
| Track | Artificial Intelligence & Agents |
| Industry | Energy & Utilities |
| Technologies | AI/BI, Unity Catalog, Agent Bricks |
| Skill Level | Intermediate |
PetroWizard & IntelliSense utilizes an agent-based Generative AI architecture on Databricks to address PETRONAS's challenge of fragmented, largely unstructured upstream data (90% unstructured, across OSDU, PI, SQL, etc.). Engineers previously spent 3-5 days locating data. This solution, featuring a Central Orchestrator and four specialized agents (Subsurface, Surface, Sensor, Model Router), enables cross-domain reasoning and model-agnostic, governed data intelligence. Outcomes include reducing data discovery time to 30–45 seconds, achieving 90%+ query accuracy, high user satisfaction (4/5 across 3,000+ users), and full audit traceability, fundamentally transforming upstream decision-making.
Session Speakers
Prashant Singh
/Sr. Solutions Architect
Databricks
Musreen Azwan B Mustafa
/Senior Manager, Surface Engineering Data
PETRONAS