AI and the Call Centre: What the British Gas Job Cuts Mean for UK Employers

Few stories crystallise the anxiety around artificial intelligence quite like a round of redundancies at a household name. In June 2026, British Gas announced plans to cut around 500 jobs across its customer operations, and within hours the headlines had a villain: the chatbot. The episode is worth reading carefully, because the disagreement at its heart (was this AI replacing people, or something more ordinary?) is exactly the conversation every UK employer will eventually have to navigate, in public.

What actually happened

According to ITV News, British Gas told staff at an online meeting that it plans to cut 500 jobs across its Services and Energy call centre teams in Glasgow, Edinburgh, Cardiff, Leicester, Stockport and Leeds, as part of a programme it calls Customer Service Transformation. The Energy team handles enquiries about gas and electricity accounts, including bills, complaints and payment difficulties; the Services team deals with breakdowns and repairs covered by HomeCare policies.

The GMB union framed it bluntly. National Secretary Charlotte Brumpton-Childs called it “an absolute disgrace,” arguing that British Gas was “slashing hundreds of human jobs and giving them to chatbots,” and that overworked, underpaid staff were “being replaced by artificial intelligence.”

British Gas disputed that characterisation directly. A spokesperson said the suggestion that “these roles are being replaced by AI or new chatbot technology is simply wrong,” pointing instead to a shift in customer behaviour. By the company’s figures, customer contacts handled by these teams have fallen from around 20 million in 2023 to 16 million in 2025, with a forecast of 14 million in 2026: an inbound drop of roughly 31% since 2023. It says about nine in ten interactions now happen through its website, app, chatbot or WhatsApp, and that it is consulting on the changes, offering voluntary redundancy where possible.

Two true things at once

It is tempting to pick a side, but the more useful reading is that both accounts can be partly true. Customers genuinely are choosing self-service channels over phone queues, and demand for live call handling really has fallen. At the same time, the digital channels absorbing that demand (chatbots, apps, automated messaging) are increasingly powered by AI. “Customers prefer self-service” and “AI is doing work people used to do” are not competing explanations so much as two descriptions of the same shift. The thing that makes this story sting is not the technology itself; it is the sense that the technology was deployed to people rather than with them.

Why it matters for businesses

For any organisation weighing automation, there are three lessons here that have nothing to do with whether the underlying technology is good.

• The narrative is part of the rollout. British Gas and the GMB are describing the same set of facts in opposite ways, and the public has largely sided with the more human framing. How you explain a change shapes how it lands with staff, customers and the press, sometimes more than the change itself.
• “Falling volume” is rarely the whole story. Even if contact really is dropping, automation tends to redistribute work rather than simply remove it: the easy queries go to the bot, and the hard, emotionally charged ones still land with a person. Plan for that mix, or the people who remain end up overloaded.
• Headcount is the bluntest possible instrument. Framing automation purely as a cost cut leaves all of the upside on the table: the same capability that deflects routine contact can also be pointed at retention, complaints resolution and faster repair scheduling: work that grows the business rather than shrinking it.

What to do about it

If your organisation is looking at AI in customer operations, the British Gas episode is a reminder to lead with the design, not the announcement. Start by mapping where routine, repetitive contact actually sits, and let automation absorb that volume first (bill explanations, appointment booking, status updates) so the gain is obvious and measurable. Be explicit with your team about what the system will and will not handle, and about where their work is moving, rather than leaving people to assume the worst from a leaked slide.

Above all, decide early whether automation is a story about doing less with fewer people or doing more with the same ones. The technology is neutral; the framing is a choice. Companies that treat AI as a way to free skilled staff from drudgery (and say so plainly) tend to keep both their people and their reputation. Those that treat it as a quiet headcount lever tend to end up in a headline they did not write.

Source: ITV News: Union says workers ‘being replaced by AI’ as British Gas cuts 500 jobs.