Most of us, at some point, have probably raised our voice to a chatbot or got frustrated repeating ourselves to an automated phone system. It usually happens when the stakes are high or when something important has gone wrong, and the last thing we want is a loop of scripted prompts. When someone reaches out to a brand, they’re rarely looking for a technical interaction. They want to feel heard, understood and supported as a human.
Recently, I heard two stories that captured this difference very clearly. An airline cancelled a customer’s return flight simply because the outbound leg wasn’t used. The response was a flat, “that’s the process.” Around the same time, a retailer chose to replace a faulty buggy well outside its warranty and said, “in this instance, we can send you a new one.” Two everyday moments. Two familiar brands. Two processes interpreted in completely different ways. What changed the outcome was not the product or the channel, but the way context was taken into account, and personal discretion used, by the person trying to help. Basically, being a human being.
AI has created remarkable opportunities to remove friction from customer journeys. It can shorten queues, simplify routine tasks and give people faster access to basic information. Yet when support slips into impersonal, unreachable or robotic – “computer says no!” – territory, trust erodes quickly.
Gartner’s recent findings reflect this tension. Only a third of customers who last spoke to a phone agent say they are ready to switch to AI alone. At the same time, almost all customer service leaders still rely on human advisers to shape and supervise AI.
This balance resonates with me. I believe the next era of customer experience belongs to organisations that use AI to lighten the load for both customers and employees, while protecting the empathy and discretion that only people bring.
At Capita, our focus is clear: use AI to speed up simple journeys and free human advisers to handle the moments that carry emotion, nuance or complexity. When we do this well, every interaction feels supported by both intelligence and care. When people and technology work together, service becomes more responsive, more intuitive and more accessible.
Human connection where it matters most
There are moments in customer service where judgement, reassurance or even compassion matter more than speed. A family navigating financial hardship. A patient seeking clarity on a health-related claim. Someone worried about fraud on their account. In these situations, a blocked escalation or a bot that repeats the same question can heighten anxiety and damage trust.
Legislation may yet reflect this. Gartner has predicted that, by 2028, the EU will enforce a “the right to talk to a human” in customer service for certain sectors. It will require organisations to give customers a direct route to a trained adviser, especially those who may be vulnerable. I would welcome this shift. It encourages responsible design, promotes fairness and reminds us that automation must serve people, not replace them.
We see this often in the work we do. A major retailer we partnered with was facing rising returns complaints and falling satisfaction. The instinct was to add more automation, but upon investigation we discovered the real issue came from an unclear policy and advisers who weren’t empowered to use judgement. Once the process was redesigned and teams were trusted to make the right call, complaints fell by more than a fifth and customer satisfaction increased significantly, with a 15-point NPS bump. AI wasn’t the answer there. A clearer process and empowered people were.
Hybrid models are becoming the new norm. They place humans at the centre, supported by transparent and explainable technology. They align with the EU AI Act’s risk-based principles and the ICO’s guidance on fairness, ensuring that oversight, judgement and empathy remain central features of customer service. When customers know they can speak to a person when it matters, their openness to AI increases naturally.
Good AI needs a strong data foundation
AI’s value depends entirely on the strength of the data beneath it. When systems are fragmented or information is poorly connected, customers feel it immediately. They repeat details, agents search through multiple screens and journeys breakdown.
When data is clean, joined up and consistent across channels, AI becomes genuinely useful. It can recognise who a customer is, understand their recent intent and offer them a seamless transition from one channel to another. It avoids unnecessary questions, supports better routing and helps advisers get to the heart of an issue quickly.
A strong data foundation gives AI the context it needs to enhance the experience rather than complicate it. At its best, this looks like a unified customer context layer that carries identity, last actions, entitlements and relevant history through the entire interaction. With this in place, AI asks less and resolves more, while respecting the customer’s time.
Enhancing adviser performance with technology
The relationship between human advisers and AI is often misunderstood. AI does not need to replace the frontline. It can strengthen it. When advisers have real-time access to accurate information, policy guidance or suggested next steps, their confidence grows. Their decision-making becomes quicker and more consistent, and they can focus on building rapport rather than navigating complex systems. By leveraging the advanced capabilities of our hyperscaler partners, we’re able to equip advisers with the latest AI tools and ensure consistent, high-quality support across regions.
We’ve seen this dynamic at scale in our work with VMO2. By equipping advisers with multilingual AI tools, a virtual call centre environment and consistent processes across regions, we helped them handle calls more quickly and deliver more consistent support. Total call handling times improved by almost a third, and customers felt the difference in the speed and clarity of conversations. These gains came from people being better supported by the process and the technology behind them.
Research from McKinsey reinforces this point. When AI sits alongside advisers, performance improves, service becomes more consistent and employees feel more capable. Customers feel that difference.
Guardrails build trust and keep organisations ahead
AI must be deployed with care. It requires thoughtful design, transparent communication and a commitment to protecting people. At Capita, we follow clear principles in how we design, build and use AI across our services. We are committed to leading with integrity, and embedding ethical, transparent and secure AI into everything we do, so that every client, colleague and community can trust the value AI delivers.
These principles help us navigate the expanding regulatory landscape, from the EU AI Act to evolving GPAI guidance. They give our clients confidence that we are not only compliant but ahead of the curve.
The path forward
The next generation of customer experience will not be defined by AI alone. It will be defined by the thoughtful partnership between people and technology. When organisations protect human connection, invest in strong data foundations and use AI with intention, service becomes more humane and more effective.
AI should remove friction from the journey. It should never remove humanity from the experience.
