Yorkshire Building Society is deploying AI agents to cut admin work in customer relations, with tools that draft complaint responses, search policies and summarise cases - saving up to 26 minutes per interaction. The mutual, which serves over three million members, unveiled the agents at an internal Data & AI festival in Bradford, showing how AI could lift the administrative burden while keeping human judgment central.
From admin overload to member conversations
"A typical day without the use of AI within customer relations can be very, very complex," said Polly Conner, senior manager of customer relations at YBS. "There's substantial regulation required within complaints handling, which means there are a lot of admin tasks to do." Switching between multiple systems, searching for policies and drafting detailed responses often delay the moment a colleague can focus on the actual conversation.
Simone Fox, Director of Customer Support at YBS, said the technology's role is to give members more choice. "This technology is going to give our members more choice and better access to information when they need it most," she said, "but there will always be a human available to speak to them when they need it."
Meet Penelope, Sam and Alf
Three agents now handle distinct parts of the complaints journey. Sam summarises long or detailed complaints, saving an estimated seven minutes per use. Penelope drafts final response letters for more complex cases, a task that can take up to 26 minutes when done manually. Alf supports both by searching policies, procedures and past cases. These AI Agents & Automation tools are embedded in a highly regulated financial environment, targeting consistency and clarity, not just speed.
"We're able to minimise the amount of time staff are spending on admin tasks and actually spend that time talking to our customers, helping them resolve the challenges they've faced," Conner said.
A single view of the member
YBS is rolling out a customer service platform built on Microsoft Dynamics 365 Contact-Centre-As-A-Service. It brings customer history, previous self-service activity and relevant guidance into one screen. An AI assistant can summarise cases, highlight knowledge articles and help draft responses - eliminating the need to hop between platforms. The society's approach to consolidating these sources reflects how AI for Customer Support can reduce context-switching and let agents resolve issues faster.
"They've got information at their fingertips," Fox added, "which means they can help members in a far more timely fashion."
Skills, confidence and control
The society invested in its data foundations beforehand, modernising its tech estate with a cloud-native data platform, strengthening data governance and tightening security. A Data and AI Academy, launched in April 2024, has seen 634 colleagues complete more than 1,500 learning modules, with 33 completing data apprenticeships - over 75% of them earning a distinction.
Fraser Ingram, Chief Operating Officer, said the organisation now has "the right foundations, the right motivation, and a good understanding of the risks and opportunities." Pete Balmforth, senior manager of operational risk controls, likened the AI assistant to "an intelligent graduate with broad knowledge," adding, "it's powerful, but it's not a substitute for judgement."
Why this matters for customer support professionals
YBS's experience shows that AI agents can absorb the most tedious parts of complaint handling - summarising histories, searching policies and drafting final letters - without removing the need for human oversight. Frontline staff gain time for empathy-driven conversations, but the governing principle remains clear: professional judgement and transparent human control are not optional, especially in regulated sectors. Teams adopting similar tools should pair technology rollouts with deliberate investment in skills and governance, so staff feel confident using AI to support decisions, not replace them.
Your membership also unlocks:
