AI Agents and Analytics Platforms for Future-Ready Utilities

Abstract: The article shows how integrating AI Agents with traditional Analytics Platforms may drive significant efficiency, expand discovery and deliver exceptional value to the Utility.

3,600 RPM. That’s the rotational speed required of a high-speed turbine coupled to a 2-pole generator to maintain 60HZ frequency (PJM Interconnection, 2018, p. 21). Lately, it seems innovations in the Utility and Analytics landscapes are spinning just as fast.

Seven years ago, it was a different story. In 2019, Electric Utilities in the US were concluding a 12-year period of essentially flat load growth (U.S. Energy Information Administration [EIA], n.d.). In that same year, Analytics were code-heavy and mainly focused on Business Intelligence – reporting on what had happened.

Fast forward to 2026 and it’s hard to keep up with everything that’s going on. Utilities are seeing record levels of electric demand. While demand grew by just 0.1% annually between 2005 and 2019, PJM is now forecasting summer peak load growth of 5.4% annually in the Dominion zone over the next decade (PJM Interconnection, 2026a, p. 34), mainly due to data center proliferation (PJM Interconnection, 2026b, pp. 33–37). Analytics have experienced the same seismic shift. Intelligent Decision and Agentic Agents seems to dominate the headlines, and AI appears to be on every executive’s wish list.

Keeping up with the speed of innovation is difficult. Calls from management to harness the latest and greatest tech trends can add to existing demands, but there’s only so much time in the day to understand, let alone master, each new analytic advancement. Thankfully we’ve arrived at a point where differing technologies are converging in a manner that may allow data scientists and business analysts to leverage the power of analytics in a more efficient and productive way than before.

Combining AI Agents with Analytics Platforms

Coupling a Large Language Model like Claude or MS Copilot with an established Analytic Platform can be a game changer that drives efficiency, productivity, and overall benefit to the Utility. The marriage of vetted analytics technology with innovative LLMs can be equated to a Combined Cycle Power Plant. The Steam Turbine, the backbone of generation, is like the traditional analytics platform, providing stability and reliable outputs. The more flexible Gas Turbine can be spun up quickly to respond to new demand – like the LLM. Together, the convergence of the Steam and Gas Turbines drive the Combined Cycle Power Plant, providing the Utility with efficiency gains, reliable power, and scalability. The combination of an analytics platform and LLM can do the same from an analytic standpoint

Analytics Platforms as Generation Assets

The Analytic Platform acts as the ‘Generation Assets’ where the heavy lifting is done. This is where your Analytics Fuel, Data, is accessed and cleansed, data pipelines are developed, models are built, results are scored, and visualizations are generated.

The LLM acts as the ‘Control Room’ telling “Generation’ what needs to be done. These instructions are delivered through the MCP Server, which acts as the line of communication.

With an LLM instructing the Analytic Platform, data scientists or analysts can potentially develop, test, and operationalize models more efficiently, simply through conversational interaction. For instance, posing questions like, ‘Tell me what the forecasted load is for the next three months.” or ‘Show me which customers are at highest risk of not paying their bill this month.” could start an analytic process and deliver results quickly.

AI Agents and Analytics Platforms: Better Together

Integrating an LLM with a traditional analytics platform can provide value far beyond the sum of its parts. In this case, each technology may help to compensate for the others’ weaknesses, providing a combined benefit that neither can achieve alone. At this point, this hybrid approach may a better way to provide defensible and explainable insights at scale. A governed analytics platform can provide the data lineage and governance that Regulated Utilities require where every data transformation is traceable, every model versioned and monitored, and every decision explainable. Most importantly, it can deliver results through proven, validated methods, and repeatable math. Meanwhile, the LLM provides a power acceleration layer to the process. Learning curves may be reduced and speed to value increased through conversational interaction along with the ability to automate documentation, summarize complex model output, and answer questions quickly. For a data scientist at a Regulated Utility, this pairing can mean faster insight generation without sacrificing governance, compliance, or mathematical defensibility.

Dollars and Cents

Efficiency and productivity are the most obvious benefits of this marriage, allowing users to potentially become much more productive. In addition to the productivity gains, there are real financial savings that can be realized. Given most of the work is completed within the Analytics Platform, the token cost may be lower compared to doing everything in the LLM. This can result in a reduced AI token spend and improve year-over-year cost predictions for budget planning.

Finally, the ease of interfacing can unlock the full power of your analytic platform. It’s common to hear people say that they only access a fraction of the power of a given platform. With this approach you can take that Ferrari that’s been sitting in the garage and drive it like you were a Formula-1 driver.

The New Baseline

The advent of AI has coincided with a shift in the Electric Utility sector from slow or even contracting growth into an expanding, innovation-driven environment. Thankfully, as demand for electricity grows and innovation become increasingly important, analytics technology continues to evolve, delivering the tools to help utilities compete and excel. The pairing of LLMs with a proven analytical platform can create a combined‑cycle effect: governed, scalable, statistically rigorous analytics accelerated by conversational AI. This hybrid model can boost productivity, reduce token costs, and deliver defensible, faster insights for regulated utilities.

For Electric Utilities, the message is clear – pairing AI with governed analytics could be a compelling path forward. Governed Analytics, turbo-charged by AI, may become the new baseline for driving Utility innovation. By embracing this model, Utilities can harness change and support more reliable, flexible, and customer-centric organizations.

References

PJM Interconnection. (2018, October 2). Generating unit basics [Training presentation]. https://www.pjm.com/-/media/DotCom/training/nerc-certifications/gen-exam-materials-nov-22-2022/training-material/03-power-system-elements/generating-unit-basics.pdf

PJM Interconnection. (2026a, January 14). 2026 PJM load forecast report. https://www.pjm.com/-/media/DotCom/library/reports-notices/load-forecast/2026-load-report.pdf

PJM Interconnection. (2026b, February). 2026 long-term load forecast supplement. https://www.pjm.com/-/media/DotCom/planning/res-adeq/load-forecast/load-forecast-supplement-2026.pdf

U.S. Energy Information Administration. (n.d.). Table 7.1. Electricity overview [Data set]. Monthly Energy Review. Retrieved September 11, 2026, from https://www.eia.gov/totalenergy/data/browser/index.php?tbl=T07.01

About the Author

Jim Magnanini serves as Director of Sales for SAS Institute’s North American Energy Practice. With more than two decades of experience partnering with Utilities, he has helped address critical industry challenges ranging from load forecasting and predictive maintenance to customer care. Jim holds a BA in English from University of Delaware and MBA in Marketing from Rutgers University.