In this blog, we’ll describe:
- Why businesses need to start with the foundations when looking towards AI.
- The importance of supporting agents first.
- How pacing out an AI rollout optimises its value.
Everyone’s talking about AI in customer experience. But the truth is that getting it right isn’t about how quickly you roll it out. What matters is how clearly it supports the people delivering the service.
Start with the foundations
Some of the loudest voices in the industry are pushing for ‘AI-first’ strategies, but most customers don’t actually want a fully automated experience. In the UK, 64% of people would prefer not to deal with AI in customer service at all.
That doesn’t mean AI has no place. It means AI needs to be used wisely, with the human side of service still front and centre.
Plenty of AI tools are being rushed in with patchy results. Replacing service with automation before fixing the basics just adds another layer of frustration.
Poorly deployed bots, IVRs with no exit route, or confusing recommendations end up frustrating customers rather than helping them. 75% of customers have experienced frustrations over AI-driven customer service.
Support agents first
One of the most powerful uses of AI is to help people do their jobs better. Think about how AI can summarise calls, surface answers faster, and spot early signs of customer frustration. These are all tasks where AI can quietly lift performance without disrupting the conversation.
But it only works if the underlying systems are ready and the data is reliable.
That was the approach taken at E.ON, where the business rolled out new contact centre tools in stages, not all at once. Agents had time to get familiar with the new systems, and slowly saw the benefit in their day-to-day work. These services were then adopted naturally.
What happened here represents a performance boost for agents, rather than just a simple tech rollout.
Slow is smooth, smooth is fast
The best outcomes come from pacing the rollout to match readiness. Rather than turning on every feature at once, the most successful teams start small and build confidence. They test features properly, adjust, and make sure there’s always a human fallback when needed.
Good governance matters too. Before you roll out anything with AI, ask:
- What problem are we solving?
- Who benefits?
- What could go wrong?
If you can’t answer clearly, you’re not ready yet.
Less hype, more value
If AI is going to work in your CX strategy, it needs to improve outcomes for real people, customers and staff alike. That means using it to smooth out the everyday bumps in service, not just ticking a trend box.
The companies who get it right in 2026 are the ones who took their time, tested carefully and made sure the results spoke for themselves.
