SMECO Turns AI Into Measurable Operational Gains

Electric cooperatives face growing pressure to improve reliability, control costs and accomplish more with limited staff. Southern Maryland Electric Cooperative (SMECO), which serves approximately 180,000 meters with about 500 employees, offers a practical example of how cooperatives can approach artificial intelligence (AI) and machine learning without overcommitting resources or chasing hype.

While many cooperatives are still evaluating AI, SMECO’s experience demonstrates that targeted operational use cases can deliver measurable value today.

Start With Problems, Not Technology

Rather than launching a broad AI initiative, SMECO focused on specific operational challenges. The cooperative established an AI acceptable use policy that safeguards confidential information while enabling employees to use AI tools responsibly. The policy also makes clear that accountability remains with the user.

“We don’t want to be a solution in search of a problem,” said Ryan Edge, vice president of strategy and innovation. “Every use case has to align with our strategic plan and deliver measurable value.”

That approach has kept the organization focused on high-impact opportunities in reliability, field operations and back-office efficiency.

Use Case 1: Outage Prediction Still Requires Iteration

SMECO’s analytics team developed a predictive model that uses outage history, weather data and more than 20 operational variables to forecast potential equipment failures within a three-day window.

“Our goal is to predict failures at the device level,” said Dan Liu, director of analytics. “We retrain the model every two weeks and continuously refine accuracy.”

The model combines machine learning with a rules-based fallback to ensure results when predictive confidence is low. Although it has not yet demonstrated measurable improvements in reliability metrics such as System Average Interruption Duration Index (SAIDI) or System Average Interruption Frequency Index (SAIFI), it is helping build the foundation for future reliability gains.

SMECO’s experience underscores an important lesson: Early AI initiatives require patience. The value extends beyond immediate savings to improving operational decision-making over time.

Use Case 2: Field Asset Identification to Reduce Response Time

SMECO is also piloting a mobile solution that allows field crews to photograph damaged equipment. AI identifies the asset, connects it to GIS and inventory systems and recommends repair actions.

“The idea is to know what failed, what parts we need and which crew is closest before they even arrive,” said Liu.

If fully deployed, the tool could reduce outage duration and improve crew utilization, both of which directly affect reliability performance and operating costs.

Use Case 3: Survey Data Automation Delivers Immediate ROI

SMECO’s strongest business case to date involves automating survey card processing. Previously, employees spent multiple days manually entering paper survey data.

“Now they scan the cards and get results in about 10 minutes,” said Sylvia Welch, vice president of information technology.

The process now takes one to two hours instead of two to three days. Depending on survey volume, reducing manual effort by an estimated 16 to 24 labor hours per cycle can generate meaningful annual savings.

This low-risk, high-return application demonstrates how AI can quickly reduce operating costs while improving productivity.

Decision Framework: Build, Buy, Partner

SMECO follows a hybrid approach to AI adoption:

• Build internally when costs are manageable and expertise is available.
• Partner when speed or scale is needed.
• Pilot solutions before committing to full deployment.

“There are a lot of great partners out there,” said Welch. “If you can’t do it in-house, someone can help you get there.”

This approach enables SMECO to control costs while maintaining momentum.

Governance, Risk Management Matter

SMECO remains intentional about AI governance, especially when data is involved.

“We want to make sure we’re using it appropriately and not exposing confidential data,” Welch said.

As a result, the cooperative has prioritized internal use cases before deploying member-facing AI tools. This approach helps reduce regulatory and reputational risk while building organizational confidence.

Practical Advice for Cooperatives

SMECO’s experience offers several lessons for cooperatives beginning or expanding AI initiatives:

• Start with a focused problem. Identify a specific operational challenge with measurable impact.
• Expect iteration, not immediate results. Most AI models require refinement before consistently delivering value.
• Measure time savings first. Productivity improvements and efficiency gains are often the fastest way to demonstrate return on investment.
• Use partners strategically. A hybrid build-and-buy approach can reduce cost and risk.
• Look beyond the cooperative network for ideas. “Co-ops often look to other co-ops for answers, but some of the most innovative solutions are coming from outside the industry,” said Edge. Exploring ideas from vendors, startups and other sectors can accelerate learning and expose new approaches.
• Establish guardrails early. Define governance and AI usage policies before scaling efforts.

Where to Start

Choose a single workflow with results that can be measured within six months. Early wins help build credibility with leadership and boards.

The Bottom Line

SMECO is still early in its AI journey, but its experience demonstrates that targeted use cases can produce immediate efficiency gains while building long-term organizational capability.

For cooperative leaders, the message is simple: start small, focus on measurable outcomes and scale what works.