Case Study: Marks & Spencer achieves 90%+ intent accuracy and simplified customer service with Twilio

Marks & Spencer Keep it simple and trust your team to build a customer service solution that lasts

Marks & Spencer, the UK retailer with a broad and varied customer base, was struggling with a legacy IVR system that forced callers into a few rigid options. With roughly 10 million calls a year, the company lacked clear visibility into why customers were calling, producing poor routing and frustrated callers.

The team built a Natural Language Platform that captured over 60,000 utterances and used intent-based routing (now over 90% accurate), paired with simple, testable prompts, rapid iterations, and human-in-the-loop handling for ambiguous cases. By prototyping in production, training advisors, and trusting the team to learn from small experiments, M&S improved call classification and routing, reduced misrouted calls, and delivered a simpler, more trustworthy customer experience.

Marks & Spencer

Akash Parmar

Enterprise Architect