Why AI matters at this scale

Murphy USA is a major American retailer operating a network of over 1,700 gasoline stations, primarily located near Walmart stores, with attached convenience stores. Founded in 1996 and headquartered in El Dorado, Arkansas, the company serves millions of customers daily, competing on fuel price and convenience. Its scale—over 15,000 employees and a vast, distributed footprint—generates enormous operational data but also presents complex challenges in logistics, inventory, and margin management. For a low-margin, high-volume business, even small efficiency gains or margin improvements, when multiplied across the entire network, translate to significant bottom-line impact. AI offers the tools to find and automate these gains at a pace and precision beyond traditional methods.

Concrete AI Opportunities with ROI Framing

1. Dynamic Fuel Pricing Optimization

Fuel is a commodity with razor-thin margins, and prices are highly sensitive to hyper-local competition, demand fluctuations, and crude oil costs. A machine learning system that ingests real-time data on competitor prices, local traffic patterns, weather, and calendar events can recommend optimal price adjustments for each station. This moves pricing from a reactive, regional strategy to a proactive, per-site tactic. The ROI is direct: capturing even a fraction of a cent more per gallon across billions of gallons sold annually adds tens of millions to annual profit.

2. Predictive Inventory for Convenience Stores

Each Murphy USA convenience store must stock a mix of fast-moving snacks, beverages, and prepared foods. Stockouts mean lost sales, while overstock, especially of perishables, leads to waste. AI models can analyze historical sales data, promotional calendars, and even local event schedules to forecast demand with high accuracy for each SKU at each store. This optimizes truckloads and reduces spoilage. The ROI manifests as increased sales from better in-stock positions and a direct reduction in shrink, protecting already tight convenience margins.

3. AI-Enhanced Site Security and Safety

With thousands of forecourts operating 24/7, monitoring for safety incidents (like spills or unsafe customer behavior) and security threats is resource-intensive. Deploying computer vision at key points can automatically detect anomalies—such as a vehicle left at a pump too long, a potential slip hazard, or unattended merchandise—and alert staff. This augments human oversight, potentially reducing liability costs and theft. The ROI includes lower insurance premiums, reduced loss, and more productive staff time.

Deployment Risks for a Large, Distributed Enterprise

Implementing AI across a network as large and geographically dispersed as Murphy USA's carries unique risks. First, data integration is a monumental task: unifying data from legacy fuel management systems, point-of-sale terminals, loyalty programs, and new IoT sensors into a reliable, clean data lake is a prerequisite for any AI model. Second, change management at this scale is critical. Store managers and associates must trust and act on AI-driven recommendations (e.g., price changes, order quantities), requiring extensive training and clear communication of benefits. Third, cybersecurity and resilience become more complex. Each connected store is a potential endpoint vulnerability, and the AI systems themselves become critical infrastructure; an outage in a pricing engine could have immediate, nationwide revenue consequences. A phased, pilot-based rollout at a subset of locations is essential to mitigate these risks before a full-scale deployment.