AI has been THE business story over the last two years, so it’s understandable utility vendors with AI capabilities have hyped them.
But the utility industry isn’t like other business sectors — especially the move-fast-and-break-things tech sector. So, it’s important to understand how AI messaging might (or might not) be landing with utility decision-makers.
In rate and safety filings, for instance, utilities themselves use very specific vocabulary to describe how they aim to deploy AI in the near term. They talk about machine learning, predictive modeling, risk-model updates, spend efficiency, and data quality.
Duke Energy’s CIO, Richard Donaldson, reflected a similar reserve in a September 2025 podcast interview. He said utilities have to do “the unsexy work” first, protecting data and governing AI platforms. That’s the opposite of hype language, and it matches the practical pattern seen in narrow production deployments to date.
According to EPRI, the Open Power AI Consortium has cataloged more than 250 AI use cases across the power sector, and others are still being prioritized, co-developed, and matched to technology maturity. That seriousness of attention shows the industry has matured beyond dreaming about transformation to looking for proof.
The proof points already exist at the category level to set the right expectations. Avista reduced high-bill service calls by 27% after deploying AMI-driven customer analytics. PG&E's internal service-planning assistant saved more than 1,400 hours and $200,000 in its first month. And Donaldson said Duke compressed regulatory-response workflows from days to seconds. That level of specificity is what utilities are now looking for.
Vendors would do well to move on from broad promises and develop messaging that is outcome-first, narrow, and utility-native.
A good way to calibrate marketing language for 2026 would be:
Use the term AI, but subordinate it to outcomes. “AI-enabled outage risk reduction” will land better than “AI transformation platform.”
Lead with measurable use cases that already have sector momentum. Wildfire and vegetation analytics, inspection QA/QC, high-bill and service-planning support, regulatory-response copilots, cyber monitoring, and forecasting are credible examples.
State the prerequisites instead of hiding them. Data quality, integration complexity, and human governance aren't weaknesses to minimize. Take the lead in addressing them directly.
Genuine momentum exists for AI in bounded customer, compliance, and knowledge-work workflows. Vendors who position it as a disciplined accelerator for work utilities already prioritize will find more traction than those still selling the transformation story.
WORTH READING
“The Year Capital Stopped Believing the Hype”
by Galen Growth, January 2026
In 2025, digital health investors also stopped rewarding speculative promise and started concentrating capital around companies that could demonstrate workflow relevance, clinical evidence, interoperability, and hard ROI.
Galen Growth's analysis of the shift is worth reading because the buying psychology it describes is almost exactly what utility procurement teams are applying to AI vendors right now. Health providers, like utilities, are complex, risk-sensitive buyers, and they reject unbounded innovation narratives in favor of proof that a product works within real operational constraints. Different sector, same moment.
NEW GROWTH SPOTLIGHT
This Hart EMC case study, developed for Itron, is worth reading as an example of what proof-oriented vendor content looks like.
Hart EMC is a northeast Georgia cooperative serving 40,000 meters across challenging terrain, and the piece describes exactly how Itron's analytics application uses what could be described as AI to help detect energy theft, maintain billing accuracy, and reduce the manual investigation burden on a lean operations team.
But the term "AI" doesn't appear anywhere in the document.
Instead, we describe how automated analytics are bounded and embedded in workflows that combine “machine and human intelligence” to “improve operational efficiency and customer service” with measurable KPIs.
The operational outcomes carry the story without need to reach for a flashy headline.
To receive CONDUCTOR by email, register here.
