In his guest blog, Shaun Smith examines the real impact of AI on customer experience in 2025. Moving beyond hype and technology jargon, he explores why the brands that truly succeeded were those that used AI.
2025 was the year AI exploded across every touchpoint—from banking apps to burger joints.
But for most customers, the real story wasn’t about algorithms. It was about how brands made
them feel. Budgets poured into automation, virtual agents, and predictive analytics—often at the
expense of traditional service improvements. Yet ask customers what they remember, and it’s
rarely large language models or orchestration platforms.
They talk about the airline that rebooked them before they reached the gate. The bank that
spotted a problem and called to check they were okay. Or the retailer whose chatbot tied them
in knots and made a simple issue feel like hard slog.
In other words, some companies are betting on AI as the solution to customer service. The best
ones are betting on CX as the focus of AI. So, which of those bets actually
paid off?
Same tools, different outcomes
In 2020, many organisations used Covid as an excuse for poor service: long call‑waiting times,
reduced hours and apologetic messages that went on for months after the lockdown ended. In
2025, the excuse has changed but the behaviour hasn’t – only now it sounds like, “That’s just
how the system works,” or “The process can’t do that, you’ll have to start again.”
Take some of the high‑street banks. Many have invested heavily in digital self‑service and AI
triage, yet customers still find themselves passed between channels, repeating their story and
fighting with chatbots that refuse to hand them over to a human. Contrast that with banks such
as DBS in Asia or Commonwealth Bank in Australia, which use AI to flag vulnerable customers,
surface richer context to advisers and proactively offer help when spending patterns suggest
someone may be in difficulty. Same tools, very different outcomes.
When AI makes good brands better
Used well, AI simply disappears into the background and lets the brand’s promise shine
through. Some airlines, for example, have upgraded their virtual assistants so customers can
rebook disrupted flights, track baggage and choose compensation options in a few taps, without
waiting in a queue. AirAsia’s “Ask Bo” service now allows passengers to change flights, request
refunds and get real‑time updates from one place, which transforms what used to be a
stressful experience into something far more manageable.
In insurance and logistics, companies such as Hiscox and Loadsure are using AI to summarise
calls, read documents and fast‑track straightforward claims so that customers get decisions in
minutes rather than days. The technology is clever, but what people remember is that their claim was sorted quickly and fairly, at a moment when they were stressed and worried about what would happen next.
One of my favourite brands is First Direct the on-line bank. We have written about the brand in
our books and twenty years on, it remains a gold standard. When First Direct started it defined
its promise as ‘The bank designed to fit around you’ and differentiated through its call centres.
Today, the first point of interaction is likely to be online using AI or a chatbot and for most
transactions it works brilliantly and it delivers its promise. But if there is a problem you are
quickly routed to a live agent where the old First Direct magic still works. I recently lost my wallet
and credit cards, but what could have easily been a very stressful situation was quickly and
empathetically resolved by a brilliant agent and a customer-centric system.
When “intelligent” experiences still feel tone‑deaf
Of course, there are still plenty of examples where AI makes a mediocre experience even
worse. McDonald’s, for instance, had to shut down an AI drive‑through pilot after customers
posted viral videos of the system repeatedly mishearing orders and adding random items to
bills. Some retailers have launched “helpful” chatbots that hallucinate answers, give out wrong
policy information or, in one car dealership’s case, agree to sell a vehicle for one dollar because
the bot had been configured to accept anything the customer said.
Then there are the restaurants that have tried to automate service only to discover that robots
cannot handle even simple breakdowns in the process –a missing item for example– leaving
guests with a worse experience than if a person had been involved. In these situations, the
problem is not the AI itself; it is the mindset. The brand is using automation mainly to cut cost or
chase novelty, rather than asking whether it improves the experience and speaks to what
matters most for customers.
What the best brands are doing differently
Behind the hype, the brands that are getting CX right in 2025 are doing three simple things well.
- Starting with the experience, not the technology
They begin by asking, “What should this feel like for our customers at the moments that
matter?” and only then decide where AI can help. Retail banks, for example, are using
hyper‑personalisation not just to push more products, but to provide timely nudges, budgeting
tips and life‑event advice that make customers feel more in control of their money. - Using AI to amplify humanity, not replace it
They let AI handle the repetitive, predictable tasks – finding information, authenticating
customers, pulling up history – so that employees are free to listen, empathise and solve
problems. In some contact centres, AI now listens in the background to provide real‑time
transcripts and sentiment cues, so advisers can focus fully on the person rather than the
keyboard. This use of AI improves the experience of customers while enriching the work for employees. If society harnesses AI in this way we shall speak less about making humans extraneous and more about making them empowered. - Being honest about trade‑offs
They are transparent about how they use data and automation, and they invite customers to
choose the level of personalisation and self‑service they are comfortable with. When AI makes
a mistake – a blocked transaction, a mis‑routed parcel – they make it easy to challenge the
decision and reach a human who can put things right.
AI vs CX in 2025: the real divide
The real divide isn’t between organisations that use AI and those that don’t. It’s between those
that use it to make customers’ lives easier—and those that use it to make their own easier.
Customers won’t talk about journey orchestration or language models. They’ll talk about the
brand that fixed their issue before they noticed. The bank that made them feel safe. The airline
that looked after them when things went wrong.
If brands keep coming back to that human experience, AI will be a powerful enabler and
productivity driver. If not, it will simply become the latest excuse to cut costs. And customers
have long memories for the brands that failed them.
So, who really won in 2025? In the end, the customer always does.
