Daniel Byrne

Marathon Petroleum is laying the groundwork for AI powered technology with a transformational program designed to standardize maintenance operations. With a focus on data quality, system integration, and process consistency, this initiative is enabling scalable AI use cases across the full IPSEC lifecycle (Identify, Plan, Schedule, Execute, Close). As Dan shares, this foundational work is critical, especially in light of a recent MIT study showing that 95% of AI initiatives deliver little to no measurable value. This session explores how Marathon is building the digital and data infrastructure needed to ensure AI delivers real operational impact and how they're actively exploring AI use cases that support predictive maintenance, intelligent scheduling, and frontline decision-making.

• Unifying maintenance processes across pipelines, terminals, and logistics, creating a consistent data layer that supports AI integration and long-term operational visibility

• How AI agents are being designed to support complex planning and scheduling decisions, factoring in asset priority, workforce availability, geospatial constraints, and material readiness

• Preparing for future use cases, including predictive modelling for asset health, computer vision for leak detection, and intelligent agents that assist with troubleshooting and parts ordering

• Generating a rich dataset across the full maintenance lifecycle, enabling continuous improvement and reshaping how operational excellence is measured and achieved