There are a few universal experiences that define modern adulthood: taxes, traffic, and the stomach-tightening dread of calling your energy supplier or broadband company. These industries — utilities, telecoms, mobile networks — have perfected a kind of bureaucratic purgatory that feels specifically engineered to break the human spirit. Anyone who has ever tried to question a bill, renegotiate a tariff, or report a broadband outage knows the ritual: endless menus, 45-minute queues, four separate transfers, and a customer-service agent who can access just enough information to apologise repeatedly but never enough to actually fix anything.
The result is a customer-service culture so consistently poor that it has become folklore. We share stories of epic failures the way previous generations swapped ghost tales. Except in this version, the ghost is usually someone trying to find out why their direct debit doubled last month.
But something has shifted. For the first time in decades, there is a genuine, structural reason to believe the misery might end. AI — not the chatbots of five years ago, with their robotic politeness and inability to understand basic sentences, but the new generation of fully autonomous systems — can now do what human support teams in these sectors never quite manage: deliver complete answers.
This is the key difference. Most customer-service interactions in utilities and telecoms don’t fail because individual staff are incompetent; they fail because the systems behind them are archaic, fragmented, and almost impossible to navigate. A human agent is often staring at four dashboards, three timelines, two billing tools, and a script demanding they stick to a call-time target. They are trying to solve a 21st-century problem with 1998 software and corporate rules written by people who have never met an actual customer.
AI doesn’t suffer from any of these limitations. It can access every relevant system simultaneously. It can read contracts, tariffs, billing histories and engineer notes without hesitation or error. It can check whether your broadband is slow because of a local outage, a router fault, or a misconfigured line. It can see instantly whether your electricity bill is wrong because of an estimated meter reading or a tariff shift that didn’t propagate across the account. Most importantly, it can actually do something about it — apply credits, adjust direct debits, reschedule appointments, correct bills — all without handing you off to three different departments.
For customers, this is transformative. Imagine messaging your energy provider and asking why your bill jumped, and instead of being placed into a digital holding pen, you receive a clear explanation, a corrected amount, and a confirmation that the problem is fixed. No theatre. No escalation. No emotional damage. Just resolution.
It is no exaggeration to say that these industries are the perfect testing ground for autonomous AI. Retail already offers fast, human-like digital experiences. Beauty and fashion brands have refined online service. Travel has been steadily automating for a decade. But the utilities and telecoms sector is stuck in a time capsule. These companies operate on margins that incentivise efficiency yet deliver some of the slowest, most frustrating support experiences imaginable. If AI can succeed here, it can succeed anywhere.
There is, of course, a human dimension. Talk of AI replacing customer-service agents is inevitably met with concern about jobs. But anyone who has worked in these environments knows the truth: the jobs themselves have been broken for years. Staff are set up to fail by outdated tools, limited access to information, and pressure to follow scripts instead of solving problems. If AI can relieve people from roles defined by burnout, hostility and impossible expectations, that is not a loss — it is an upgrade.
More importantly, the public has suffered the consequences of this dysfunctional system for long enough. Every hour spent on hold, every circular conversation, every incorrect bill and cancelled appointment represents a failure of design, not a failure of the customer. AI’s arrival offers something vanishingly rare in this space: hope. The hope that the industries that invented bad service might — finally — provide something better.
For the first time in decades, the pain of dealing with energy and telecom giants might actually go away. Not because they suddenly decided to care, but because a new technology makes it impossible for them not to.
And frankly, it’s about time.
Written by Daniel Yeo
