Why AI matters at this scale

Global Partners LP is a large, established master limited partnership operating in the Northeast US, primarily engaged in the storage, distribution, and retail marketing of gasoline, distillates, residual oil, and renewable fuels. The company owns, controls, or has access to one of the largest terminal networks in the region, along with a portfolio of retail gasoline stations and convenience stores. Their business is fundamentally about the high-volume, low-margin physical movement and sale of energy commodities, where operational efficiency and supply chain reliability are paramount.

For a company of this size (1,001-5,000 employees) and vintage (founded 1933), competing requires moving beyond legacy manual processes and fragmented data systems. AI presents a critical lever to compress costs, enhance margin capture, and manage complex regulatory and safety requirements. At this mid-market scale within a capital-intensive sector, targeted AI adoption can deliver outsized ROI without the bureaucratic inertia of mega-corporations, allowing for agile piloting in high-impact areas like logistics and predictive maintenance.

Concrete AI Opportunities with ROI Framing

1. Predictive Maintenance for Terminal Assets and Fleet: Unplanned downtime at a bulk liquid terminal or in the delivery fleet is extraordinarily costly, leading to supply disruptions and emergency repair bills. By implementing AI-driven predictive maintenance using IoT sensor data from pumps, valves, and transport trucks, Global Partners can shift from reactive to condition-based upkeep. This reduces maintenance costs by 10-25%, extends asset life, and prevents environmental incidents, directly protecting revenue and minimizing regulatory fines.

2. Dynamic Margin Optimization and Pricing: Fuel margins are volatile and hyper-local. AI models can synthesize real-time data streams—including rack prices, local competitor pricing, demand patterns, and even weather forecasts—to recommend optimal wholesale and retail pricing. This dynamic pricing capability can lift fuel margins by 1-3%, a significant impact on volume, translating to millions in annual incremental EBITDA for a distributor of this scale.

3. Automated Compliance and Safety Monitoring: The company operates in a heavily regulated environment with strict requirements for environmental protection, safety, and reporting. AI can continuously monitor sensor networks across terminals for potential leaks, vapor emissions, or safety protocol deviations. It can automate the generation of compliance reports and trigger immediate alerts. This reduces manual labor, lowers the risk of human error in reporting, and provides an auditable digital trail, mitigating legal and reputational risk.

Deployment Risks Specific to This Size Band

Companies in the 1,001-5,000 employee range face unique AI deployment challenges. They possess significant operational complexity and data volume but often lack the extensive in-house data science teams and unified IT architecture of larger enterprises. Key risks include:

• Legacy System Integration: Core operations may rely on older ERP (e.g., SAP, Oracle) and operational technology systems. Extracting and cleansing real-time data from these silos for AI models is a major technical and budgetary hurdle.
• Talent Gap: Attracting and retaining AI/ML talent is difficult when competing against tech giants and pure-play digital firms. This necessitates a strategy blending strategic hiring, upskilling of operations-savvy analysts, and leveraging vendor-supported AI solutions.
• Pilot-to-Production Scaling: Successfully demonstrating an AI pilot in one terminal or region is different from rolling it out enterprise-wide. Scaling requires buy-in from multiple operational divisions, standardized data pipelines, and robust MLOps practices, which can strain existing IT resources.
• Change Management: In a traditional industry, frontline workers and managers may be skeptical of AI-driven recommendations. Effective deployment requires transparent communication, demonstrating AI as a tool to augment (not replace) expertise, and designing user-friendly interfaces integrated into existing workflows.