Automating claims conversations at scale.

HomeServe handles millions of customer calls every year across its UK home services operation. With growing call volumes and increasing pressure on operational costs, they needed a way to handle more conversations without proportionally growing headcount.

We worked with HomeServe to design and deliver a voice AI solution that could handle real customer claims conversations — not just simple routing, but meaningful, end-to-end dialogue.

---

The challenge

HomeServe's contact centre was handling over 10 million calls a year. A significant proportion of these were claims-related — conversations that followed repeatable patterns but still required careful handling to meet customer expectations and regulatory requirements.

Previous automation efforts had focused on simple IVR and call routing. The challenge was to go further: to automate the substance of the conversation itself, while maintaining the quality and empathy that customers expect.

---

The solution

We designed and built a voice AI platform capable of handling claims conversations from start to finish. The system was built to understand natural speech, manage multi-turn dialogue, and integrate with HomeServe's existing back-office systems.

We took an iterative approach — starting with a focused proof of concept, then expanding coverage as confidence in the system grew. Each sprint was informed by real call data and customer feedback.

---

The results

The voice AI platform now handles a significant volume of HomeServe's claims calls, automating 20–30% of conversations that previously required a human agent.

This has contained £8.5m in annual operational workload, allowing HomeServe to manage growing call volumes without proportional increases in headcount.

The platform continues to evolve, with new conversation types being added as the business identifies further opportunities for automation.