“The question isn’t whether AI will transform the customer journey. It’s whether we’re transforming it with intention or in desperation.”
In 2025, businesses face a dual crisis: employees fearing obsolescence as AI rolls out across customer touchpoints, whilst organisations race to deploy AI solutions they barely understand. Research shows that whilst 90% of brands have deployed or plan to deploy AI in customer experience, only 5% consider their systems mature and optimised. But the deeper problem is this: most companies are layering AI onto customer experience foundations that were already mediocre. This article explores why excellence must come before automation, and why the gap between AI adoption and genuine customer experience readiness is creating frustrated customers and anxious employees.
Two Conversations, One Silent Collision
Walk into any boardroom today and you’ll hear executives discussing deployment timelines and efficiency gains for AI. Walk onto the contact centre floor, and you’ll hear something entirely different: “Will I still have a job next year?”
We’re in the middle of a transformation where 86% of customer service professionals report testing or implementing AI solutions, yet 83% of UK consumers still prefer speaking to a real person for service.
The question everyone’s grappling with: Should we deploy AI before it’s perfect, learning in real-time with real customers? Or do we wait for the “gold standard” whilst competitors forge ahead?
The Foundation Problem: You Can’t Build Excellence on Mediocrity
But here’s the question no one wants to ask. What if the foundation you’re building on is already broken?
Most companies don’t actually know what excellent customer experience looks like. They aim for “satisfied” because that keeps contracted customers from leaving. But satisfied isn’t delight. Satisfied isn’t loyalty. Satisfied is simply the absence of enough pain to trigger the effort of switching.
This matters because AI amplifies whatever foundation exists. If your base customer experience is mediocre, AI won’t transform it. It will magnify the mediocrity at scale.
“Adding AI to poor customer experience doesn’t create transformation. It creates frustration at scale.”
Consider a recent experience shared by a colleague and friend. She’s been with a major broadband provider for eight years. Out of contract and paying £64 per month for 350Mb/s when competitors were offering 1Gb/s for £35.
She started on the company’s app, engaging with their AI bot to discuss upgrading. The bot was, in her words, “absolutely useless.” Minutes of frustration with no useful responses.
Then a human agent joined. Barely better. Several more minutes going backwards from £81 per month, explaining that competitor offers are for new customers only. Eventually, they reached £65 for a gig. “Just £1 more per month for so much more bandwidth,” the agent proudly typed.
Her response: “Don’t worry, I’ll just become a new customer elsewhere and we can disregard my eight years. Bye.”
At this point, a retention specialist joined. 1Gb/s for £28 per month. She would have paid £35.
The AI was terrible. The first human was barely better. Only the retention specialist understood customer value. This isn’t an AI problem. This is a customer experience foundation problem that AI simply amplified.
She has re-contracted, but she wouldn’t recommend or refer the company. Their base level customer service is poor. She just didn’t want to “replant the tree.”
Interestingly, she has another service with the same provider at a property in Belfast. Different team. Exceptional CX that delights. Same company, different foundation, completely different experience.
Before rushing to deploy AI, companies need to ask a harder question: Is our current human-led customer experience actually excellent? If not, layering AI on top won’t fix the problem. It will just automate mediocrity.
The Rush to Amplify What’s Broken
Against this backdrop of mediocre foundations, companies are racing to deploy AI. Whilst 90% of brands have deployed or plan to deploy AI in customer experience, only 8% have fully deployed AI at scale, and a mere 5% consider their systems mature and optimised.
Whilst 71% of UK organisations using AI report better productivity, these same studies highlight persistent concerns about customer trust, accuracy, and the risk of alienating customers.
Companies using advanced AI see up to 57% more handling capacity and cost reductions exceeding 20%. But efficiency alone does not guarantee satisfaction or loyalty. More than 30% of UK consumers will choose AI-automated interactions if it means lower prices, but they reward human interaction with trust and loyalty.
“The pandemic taught us that waiting for perfect conditions means never starting. But deploying half-ready AI onto customers can destroy trust that took years to build.”
The Real AI Challenge: Three Critical Elements
AI’s performance is only as good as the foundation you build for it. Three elements determine whether your AI delivers exceptional or exasperating experiences:
Context. Does your AI understand who the customer is, what their history looks like, and what matters to them? Without rich contextual awareness, AI defaults to generic responses that miss the mark.
Specific Goals and Instructions. Too many organisations treat AI like a Google search. Effective AI needs clear directives: What are we trying to achieve? What does success look like? What are the boundaries?
Relevant Data and Examples. Clean, structured data across systems is non-negotiable. Your AI needs to learn from real examples of excellent interactions, not just process data in a vacuum.
Most organisations still struggle with data quality and integration. The companies succeeding with AI aren’t the ones with the biggest technology budgets. They’re the ones investing heavily in governance, controls, and contextual training data.
“AI isn’t a plug-and-play solution. It’s a mirror that reflects the quality of your operational foundations back at you, magnified.”
When the Happy Path Breaks: The Gaps in the Experience
Let me share my own recent experience that crystallises this challenge.
I placed an online order. The browsing experience was seamless. Payment was frictionless. The confirmation email arrived instantly. Everything on the “happy path” worked beautifully.
Then part of my order didn’t arrive.
Suddenly, the seamless digital veneer cracked. I called customer service and found myself speaking with an Agentic AI Agent. Not a simple chatbot, but a conversational AI designed to handle phone enquiries. I explained the situation. A straightforward issue that needed resolving.
Here’s where it fell apart. The AI agent couldn’t resolve the issue. When I asked to speak to a human agent, it informed me that I would need to email customer service instead. Speaking to a human over the phone wasn’t possible. Not because the technology couldn’t facilitate it, Agentic AI certainly has that capability, but because the company had programmed it to direct callers to email rather than offering phone-based human support.
Eventually, I resorted to email. The company did respond swiftly. But the experience was anxiety-inducing. Would they reply in five minutes or five days? Was my email lost in a queue? I felt trapped in a black hole of uncertainty.
This is the gap in so many customer experience journeys today. Companies deploy AI to handle enquiries efficiently, but they deliberately programme AI to block access to human support, using it as a cost-saving gatekeeper rather than a customer enablement tool.
Here’s what customers actually need. We don’t mind self-service when it works. Getting the right information quickly to move forward is brilliant. But when we can’t find what we need, when we’re forced to hunt through FAQs or wait indefinitely for email responses, it becomes deeply frustrating.
The lesson is clear. If you use AI to create a wall between you and your customers during moments of need, you will lose them. If you use AI to empower quick resolutions whilst providing clear escalation paths to humans, you’ll keep them for life. The technology can do this. The question is whether companies choose to enable it.
“Customers don’t hate AI. They hate being trapped in experiences that don’t acknowledge AI’s limitations.”
The UK Reality: A Generational Divide
Only 13% of UK consumers believe AI will support jobs, revealing deep scepticism about AI’s impact on employment. Meanwhile, 84% of CX leaders say AI adoption is “very important,” showing a striking gap between leadership conviction and public trust.
Generational splits are emerging. UK Gen Z and Millennial consumers show much higher enthusiasm for AI-driven personalisation, with 44-55% appreciating these features compared to Boomers. Acceptance will accelerate over time, but organisations can’t assume blanket acceptance today.
More than 70% of customers find chatbots helpful for simple issues, but they expect easy escalation to human agents when issues become complex or emotionally charged.
For UK businesses, the mandate is clear: deploy AI where it genuinely improves customer experience, make human escalation seamless, and never sacrifice trust for efficiency.
The Path Forward: Foundation First, Orchestration Second
Before you race to deploy AI, ask yourself: Is your current customer experience actually excellent? Not just adequate. Genuinely excellent.
If the answer is no, you have a choice. Fix the foundation first, or accept that you’re about to automate and scale mediocrity.
The companies making real progress understand this sequence. They’re starting with clear, bounded use cases where AI can deliver immediate value, building feedback loops, maintaining easy escalation paths to human agents, and being transparent with customers about AI’s role. But critically, they’re doing this on a foundation of customer experience that already works.
This moment isn’t about choosing humans versus AI. It’s about orchestrating them effectively:
- AI where it delivers speed, consistency, and scalability for routine interactions
- Humans where they deliver empathy, judgement, and trust for complex or emotional moments
- Seamless handoffs that feel like continuity, not escalation
- Clear escape routes when AI reaches its limits
Conclusion: Excellence First, Automation Second
If you’re a CX leader today, start with an honest audit. Is your current customer experience genuinely excellent? Do your customers feel valued, or just adequately serviced? Are you creating delight and loyalty, or just preventing enough pain to avoid churn?
If the foundation isn’t solid, fix that first. Then layer AI strategically, starting with one use case. Get it right. Learn what works. Then expand. Focus obsessively on giving your AI the context, specific goals, and relevant examples it needs to perform. Make human escalation seamless. Measure not just efficiency, but customer satisfaction and loyalty.
The race isn’t to deploy AI first. It’s to deploy AI effectively, in ways that make your customers’ lives genuinely easier and your employees’ work genuinely better.
That’s not a race you win with speed. It’s a race you win with intention, built on a foundation of Operational Excellence and genuine customer care.
“The future isn’t purely AI or purely human. It’s the companies that build excellence first, then amplify it with AI, who will lead the next era of customer experience.”
References
- Customer Experience Magazine: “Survey: A customer-first approach and AI adoption drive CX improvements” (2025)
- Sobot: “AI customer service response trends and stats in 2025” (2025)
- TechRadar: “Over two-thirds of retailers have already partially deployed AI agents for efficiency” (2025)
- Business Wire: “Survey Finds Agentic AI Helps CX Teams Handle 57% More Customer Service Tickets” (2025)
- Capterra: “Nearly half of UK Contact Centre Teams Use AI-Powered Software” (2025)
- 8×8/Streetview: “UK Customers Still Prefer Human Support Over AI” (2025)
- KPMG UK: “Despite the hype, there’s a long way to go to convince consumers of the benefits of AI” (2023)
- Statista: “Consumer sentiment on AI personalisation in the UK 2024” (2024)
If you’re about to layer AI onto a customer experience that is merely adequate, the foundation is the place to start, not the automation. To talk through how to get the operational groundwork right before you scale AI across your customer journey, get in touch.