The agents have already clocked in: AI’s next shift in utility customer operations
AI capability in utility customer operations is progressing across three distinct horizons, and where a supplier sits on that progression matters more than any single deployment decision today. Customer operations is also where GenAI is finding its natural foothold. While traditional AI investments in utilities have often focused on grids and assets, GenAI excels where language, context, personalization, and customer engagement intersect. This explains why customer service, collections, onboarding, revenue assurance, and customer communications are rapidly becoming the next frontier for AI adoption across European utilities.
Horizon 1 is assisting AI: copilot-style suggestions, summaries, and knowledge search that keep a human firmly in the driver’s seat while supporting day-to-day decision-making and task execution. Many suppliers are already leveraging Horizon 1 capabilities embedded in leading enterprise platforms.
In practice, Horizon 1 is already visible across customer operations through AI-assisted case summaries, knowledge navigation, bill-shock prevention alerts, and copilots that help employees resolve customer queries faster and more consistently.
Horizon 2 is agentic AI, where systems move from summarizing information to acting on it, executing defined tasks with human oversight rather than waiting to be asked.
Horizon 3 is autonomous AI: self-routing, self-healing processes where humans shift toward review and exception handling. Most European suppliers are active in Horizon 1 and beginning to enter Horizon 2 today, while Horizon 3 remains aspirational for now, since fully autonomous decision-making in processes that touch customer trust and revenue directly is not something suppliers will hand over lightly.
It is Horizon 2 where the sharpest near-term value is concentrating. Debt management is the clearest example today, though far from the only one. Revenue assurance, fraud and hardship differentiation, customer management, and market communication processes are also emerging as strong candidates for agentic AI. IDC’s Energy Consumer Affordability and Trust Survey found that more than a third of energy customers fell behind on payments or paid late in the past 12 months, and the pattern is slightly worse among Gen Z and millennial customers, a growing share of the base.
That gap is modest today but trending in the wrong direction, and it compounds: the margin exposure suppliers are managing now is smaller than the one they will be managing in five years if collections processes do not change. IDC predicts that by 2026, 45% of energy suppliers will use agentic AI for payment planning and debt management, increasing recovery rates by 20% while also boosting customer trust and satisfaction.
The mechanics matter. Agentic AI does not wait for a collections agent to flag a problem account, it identifies customers drifting into arrears early, scores payment propensity, and proactively engages before debt escalates, all of which links directly to unbilled revenue reduction and faster cash collection. Layered onto this is a capability suppliers have historically underinvested in: distinguishing genuine financial hardship from fraudulent nonpayment.
Framed correctly, this is not about policing bad actors, it is about protecting every paying customer from absorbing the cost of those who are not. It also reframes collections from a cost center suppliers tolerate into a trust-building function customers actually notice; the difference between a threatening letter and a proactive payment plan before an account ever reaches arrears is the difference between a customer who churns and one who stays.
Two further forces explain why this progression is accelerating now.
The first is electrification, which is layering real complexity onto what suppliers must sell, price, and manage, as EVs, DERs, dynamic tariffs, and bundled offerings become mainstream. The commercial stakes are real: solar panels, home battery storage, EV charging points, heat pumps, and smart thermostats are all part of this mix, but only 39% of customers own three or more of them, the threshold at which satisfaction and investment appetite compound into a self-reinforcing cycle. The bulk of the opportunity still sits with customers moving through their first one or two technologies, a segment that rewards suppliers who can sequence and launch bundles quickly, reaching customers before competitors do. This is itself another Horizon 2 opportunity: agentic AI can assemble, price, and test bundles at a speed no manual tariff team can match, then match them to the customers most likely to adopt.
The second force is a decade of point solutions. European suppliers are integration rich and insight constrained, having layered CIS, CRM, billing, and DER systems on top of one another without making the data inside them easy or affordable to act on. Only 7.5% have fully rationalized their application portfolio in the cloud, and just 16% have reached enterprise-level AI integration. None of the Horizon 2 use cases above perform well on fragmented, poorly governed data, and GenAI is now forcing suppliers to confront that data estate directly, a clean-up considerably easier inside a unified platform than across a patchwork of point solutions.
There is considerably more to this picture. IDC Energy Insights will explore how utilities are moving through distinct AI horizons, from today’s embedded copilot capabilities to tomorrow’s agentic and, eventually, autonomous operating models.
The InfoBrief will examine the practical use cases already delivering value, the role of clean data and modern platforms, and why the platform decisions utilities make today will ultimately determine how quickly they can capitalize on the next generation of AI innovation. For utilities evaluating their next AI investment, understanding where today’s copilots end and tomorrow’s agents begin will become increasingly important.
You’ve just had an exclusive preview of the insights to come. This preview introduces some of the key trends shaping the future of AI in utility customer operations, but there’s much more to discover. In September, IDC will publish the full InfoBrief, offering deeper market insights, practical use cases, and strategic recommendations to help utilities unlock the full potential of AI.
Register today to be among the first to receive the complete InfoBrief directly in your inbox as soon as it’s released. Don’t miss out on this opportunity to access valuable IDC research: completely free of charge.
Fill in the form below to secure your copy and stay ahead of the latest developments in AI for utilities.