DataRobot and Chevron Launch Joint Project to Advance Autonomous Inspections with Agentic AI

DataRobot, a provider of an Agentic Workforce platform, today announced a collaboration with Chevron U.S.A. Inc., a subsidiary of Chevron Corporation, to apply agentic AI to edge environments, supporting autonomous patrol and inspection tasks at Chevron's facilities.

This project supports Chevron's 'Facilities and Operations of the Future' initiative, focusing on using AI agents to improve how missions for robots patrolling the site are planned, evaluated, and executed within established operational standards.

Chevron uses aerial drones and ground robots to patrol and monitor equipment anomalies worldwide. Previously, for every mission, operators had to manually verify the robot's status through an authorization process. This introduced additional manual steps and confirmation time before inspection work could begin.

Debanjan Saha, CEO of DataRobot, stated: 'Agentic AI allows for a shift from a style where every mission must be manually approved to one where evaluation is real-time and continuous. Using DataRobot's Agentic Workforce platform, Chevron can deploy AI-driven decision-making at the edge while maintaining security, reliability, and governance requirements.'

Chevron is leveraging DataRobot to generate mission plans using advanced optimization and inference agents, utilizing NVIDIA foundation AI software and compute power within Chevron's digital and operational systems. Unlike point solutions built for single tasks, DataRobot's Agentic Workforce platform integrates specialized agents—from sensor analysis to geospatial inference models—into a cohesive workflow. This allows integration with Chevron’s existing systems without the need for infrastructure overhauls, enabling teams to evaluate advanced AI capabilities closer to the operational site.

This approach applies evaluation capabilities through a 'Safe Start' agent, realized via NVIDIA Inference Microservices (NIMS) within the DataRobot platform, to assess the operational status before and during robot missions. The system utilizes existing wired gas sensors, complementary vision systems, and AI models to assess site conditions.

Cari Armpriester, program manager for Facilities and Operations of the Future at Chevron, noted: 'Simply put, we are reducing reliance on manual point-in-time verification and continuously validating conditions during operations, which improves safety. This allows us to reduce exposure risks, consistently apply safety standards, and deploy robots more efficiently across our operations.'

This joint project shifts the focus from 'robot hardware certification' to 'environmental condition assurance,' helping to deploy autonomous inspections more consistently while maintaining oversight and control with reduced human intervention.