AI in customer service is the safe bet, going further holds the real upside

It's time to talk about how UK businesses are deploying AI

Launching in a new market, you come up against the prevailing conditions. How are companies experiencing the problem your product solves, and do their established habits help or hinder them in seeing a better way? Nebuly is a broad church by design; we analyse any agentic conversational product in almost any sector, with clients from farming equipment manufacturers to luxury car makers, health tech and banks in between. A few hundred conversations later, I wanted to share what factors we see holding some UK businesses back.

This is the first of three short pieces. While plenty has been said about the challenges getting from pilot-to-production and the need for governance, all of it valid; I want to explore a more structural perspective; that putting AI in your customer support function or contact centre is simultaneously the safest bet and the most limited application of the technology.

The first part below will unpack this view, before highlighting companies going further - a sort of Hall of Fame. Part two goes deeper on our client Oura as a standout leader in merging the worlds of customer support, CX and AI; and part three will offer some starter thoughts on how this ties to fundamental questions of competitive advantage and brand strategy.

Part one: thirty years of learning to listen

Over three decades the contact centre has come to dominate how UK businesses invest in serving & understanding their customers. Hardly surprising in an economy where services have grown from around 70% of output to more than 80%. Alongside this trend, marketing evolved from telling to listening as aspirational, product-led gave way to insights driven, consumer-led messaging. With digital and mobile the discipline continued to reorganise around consumer insight, and I watched the tail end of that at Qualtrics, working with names like Samsung, Unilever and Vodafone. Having the best product was no longer enough, "Experience" became the thing that differentiated your brand.

An enormous software category grew up to serve the shift: Qualtrics, Medallia, Sprinklr, NICE, ServiceNow, InMoment, Intercom, now Fin (so many more) plus the smaller companies with one valuable capability acquired and bolted on. All of them promising to connect insights to action in one way or another and drive better CX. An honest caveat, there are obviously differences in how these providers and others in the space emphasize the way they drive value – from contact centre optimisation to more granular experiential insights for product, brand etc. – but really, it’s the same game in my opinion.

An ironic twist few saw coming

A movement founded on the premise that customer experience is felt, subjective and qualitative ended up encoding itself almost entirely in hard metrics. A score can be owned, targeted and improved, a description of how someone felt, not so much. Sentiment analysis doesn't fully fix this, those signals get pressed into the same cost game. No single vendor is to blame either; it's the problem with category building. Efforts to build an ever-better system-of-action compound, and action trumps insight in terms of operational priority. Nearly thirty years on, CX measures operational performance more closely than it measures customers' actual experience.

That's not to say it isn't valuable. I've seen what Qualtrics and its peers deliver in that context, and I know the limitations too. But the value comes from optimising an operation you can already see: your channels, your queues, your teams. Contact centre metrics are a lossy summary of experience. That’s fine when journeys are predictable enough that one number adequately represents thousands of near-identical interactions, and when there are convenient moments to ask "how would you rate your experience?".

But almost none of it carries over to AI as a channel. A good AI experience isn't navigation, it's delegation, and ideally for your customer it's end-to-end. Interrupt them to ask how it's going and you break the magic. Once every conversation is its own journey, an aggregate score is less representative. So, teams fall back on adoption and token counts, which sit downstream of experience quality: they'll tell you something went wrong, but not what.

Worse, many think this is all they need, with a little extra focus on “outcomes”, as measured by, drumroll… a survey. If the only tool you have is a hammer, everything looks like a nail.

What that conditioning does to your support use case and your AI horizon

Buyers haven’t only built up a toolkit over these three decades, they’ve built a preference for evidence that arrives with a number attached. Business cases get funded when they do, and the cleanest numbers are found in the contact centre. Drop AI in there and the case writes itself.

Which is how a cost metric ends up passing as a CX metric. CSAT at least has an experience component. Containment doesn't pretend to: it counts throughput and gets reported as if it tells you something about how customers were served. Hence strategies like Deflection and Containment, terms you'd sooner associate with a zombie horde than with your paying customers.

Take two companies for example, one claiming 50% of cases handled by its AI agent and another claiming 80%, you'd be forgiven for assuming the second is doing better, as though it’s a useful way of ranking companies service. But that 80% figure is equally consistent with a good agent papering over a bad service, and a bad agent trapping people who'd have been better off speaking to a human. The number counts throughput and too often teams aren’t equipped to differentiate these outcomes. The answer lives inside the conversations but the inherited kit was built to measure something else.

As a result, the familiar experience follows the same pattern you’ll see on Trustpilot and Reddit: doom loops, no route to a human, and "I'm sorry, I don't have that information, please check our website"; all persistent signs of poor AI products serving as the face of your brand.

So, AI customer support agents: help containment metrics, can hurt customer experience if too generic, and don’t help you differentiate your brand. For this you need a set of other tools on top of your contact centre & experience insights kit.

What you forfeit by stopping at the support use case

• A richer source of insight. There's a limit on what support can tell you, however well you instrument it. Customers turn up because something went wrong, which makes it a good record of friction and a poor one of what people want when nothing has. Conversational AI is the most direct read on your customers you've ever had access to; it seems a waste to point it only at the complaints.
  
  

• The differentiated product you never build. Building on this more generalised source of insights to design dynamically personalised experiences that are revenue generating (agentic commerce).
  
  

• An owned surface to defend from. If your customers ask a general assistant about your products, it will answer, using whatever it has picked up, and you have no say in what it says. Your own conversational agent means safeguarding your brand values and customer relationships through curated, trusted experiences.

We will explore these points in more detail in Parts 2 & 3.

The Hall of Fame

Nobody’s perfect, but some are trying harder than others, and that’s important.

Revolut runs two agents, deliberately. Their support AI handles around 75% of questions with 2,000-plus human agents behind it, and they shipped AIR in April anyway: a separate in-app assistant that reads spending patterns, manages cards and subscriptions, tracks investments and plans travel, now live in France, Ireland and Spain too.

The distinction matters because AIR isn't there to reduce the cost of a problem. It's there when nothing has gone wrong. Ask it to sort out connectivity for a trip and it will find the eSIM plans for your destination and sell you one in the chat. That's a revenue surface, an owned relationship, and a record of what customers want rather than what they couldn't get, out of one product Revolut chose to build separately from its help desk.

Virgin Atlantic already had a contact centre agent. Amelia has been answering common questions in the help centre for years, and when they built the AI Concierge in December 2025 they didn't upgrade her, they started from the most valuable conversation in the business instead of the cheapest. Their best holiday baskets came from retail stores, so the team watched those reps and boiled it down to the seven things a good one finds out before recommending anything. Speaking at AI Summit London in June, VP Data & AI Richard Masters put it at roughly half a million conversations and about £1m of revenue, and mentioned in passing that the Concierge has improved their other channels too. On brand, his line was that you can't put a human in someone's living room at midnight but you can put something of that essence into a chatbot, hence a concierge voice seeded from recordings of their own cabin crew. That's brand distinctiveness as an engineering practice rather than a value on a wall.

And finally, not a UK company, but a Nebuly client, Oura: Read part 2.

Nebuly

Nebuly is the ROI platform for enterprise AI. It connects to the AI agents your business runs on, the assistants your customers interact with, and the tools your employees use every day, including Claude, ChatGPT, and Copilot, and translates that activity into business value. How much time is being saved across teams. What revenue your AI is influencing. What adoption and AI proficiency look like in practice, across departments and geographies. If you need clarity on what your AI investment is actually delivering, book a demo.