BCG: AI Could Cut Energy Retailers' Service Costs 40 to 60 Percent

Large language model customer service could cut an energy retailer’s contact center, collections, billing and back-office costs by 40 to 60 percent over three to five years, BCG estimates — and the same technology is about to make those customers much easier to lose. The analysis comes from BCG authors including Christian Wagener and Jochen Weber, in BCG’s published report.

BCG says AI changes retail energy economics in three ways. First, cost: the 40 to 60 percent service reduction. Second, switching: AI comparison platforms and personal switching agents lower the effort, and new EU rules effective by year-end 2026 require suppliers to complete a switch within 24 hours, which threatens recovery of acquisition costs of 80 to 150 euros per European household. Third, inertia: EU switching stood at just 9 percent in 2024, which BCG attributes to friction that is disappearing.

Its answer is to compete on a household’s total energy bill instead of per-kilowatt-hour margin. Integrated management of power, heating and cooling, and electrified transport could cut a typical annual bill of 3,500 to 5,000 euros to about 1,400 euros today, or 700 under optimal conditions, a 60 to 80 percent reduction, with self-generation accounting for roughly half of the savings. Oil majors, automakers, telcos and tech entrants are already moving in.

BCG recommends modeling the total bill, deciding whether to own the customer interface, and securing installer and platform partners. Where no new model exists yet, such as North America, it recommends applying AI to the existing book of customers to cut acquisition and service costs.

The lesson reaches beyond energy. When AI lowers the cost to serve and the cost to switch at the same time, customer inertia stops protecting margin, and the outcome that matters is what the company offers beyond the unit price.