Embedding AI without losing the human touch

AI is rapidly moving up the agenda for building societies. Three‑quarters of UK financial institutions were using AI as early as 2024, according to the Bank of England, and that number is only accelerating. The debate now is not whether building societies should adopt AI, but how it can be used to genuinely improve performance and member outcomes.

At PA Consulting, our research shows that members of building societies remain cautious in their attitudes towards AI. Only 35% of members feel comfortable with AI’s use in financial services, and just 29% say they understand how it is used, leaving the majority uncertain or unaware. Customers across the sector voiced concerns about accountability, data privacy, and losing human support. They wanted to know who would be responsible if AI offered poor financial advice, or worried that it wasn’t nuanced enough to handle complex problems.

This raises a critical question: how do building societies embrace AI’s potential without seeming to lose the trust and human connection that sit at the heart of the mutual model?

1. Real value – not gimmicks

First, AI adoption must remain purpose-driven. Customers are quick to sense when AI is being used for novelty rather than need, or when it complicates interactions instead of simplifying them. Real value means applying AI to areas where it genuinely improves outcomes.

For example, at Nationwide, we are piloting tools that take repetitive work off colleagues’ plates, such as conducting quality assessments, and delivering real-time colleague AI assistance to better serve our customers. Rather than using AI to replace people, this actually frees colleagues up to spend more time on the empathetic, human interactions that members value most. In other words, the technology strategy is designed to reinforce our human-centred values.

Another example is in supporting vulnerable members. Over a third of Nationwide’s interactions are with vulnerable customers, many of whom face multiple vulnerabilities. AI can play a critical role in supporting these people and ensuring their future needs are met. For example, within our chat teams, if a customer expresses a concern about their physical safety, AI tools push a real-time alert to the team leader to intervene and help manage the situation appropriately.

AI even has the capacity to analyse members’ tone of voice to flag distress or confusion, which could in future prompt frontline colleagues to dig deeper into early indicators of vulnerability, such as the mention of a spouse in hospital or impending redundancy. And it can role-play customers with complex or vulnerable needs during staff training, enabling staff to practice handling tough situations with confidence and empathy.

Applications like these aren’t flashy, but behind the scenes, they can quietly enhance the quality of member interaction. Rather than replacing human care, AI can cut through the administrative noise and help colleagues be even more present and thoughtful with members.

2. Building member confidence

Second, communication about AI governance and use is just as important as the technology itself. For example, Nationwide follows an enhanced governance process with specialist review, challenge and oversight for developing and implementing Generative AI technologies, ensuring member outcomes are protected.

Building societies must ultimately make AI understandable, human, and clearly beneficial. From our combined experience across the Financial Ombudsman Service, retail banking, and customer services, we have seen first-hand that people don’t want technical explanations of algorithms. They want straightforward answers about what the technology does, why it’s used, and how much control they have.

Even explanations like ‘we use assistive technology to give you quicker, clearer answers and keep your money safe’ are clearer than ‘we use AI bots to process your request,’ reinforcing that the member – not the machine – stays in control.

This also means adopting AI in a way that builds members’ confidence rather than forcing it upon them. As part of this, Nationwide focuses on offering customers their channel of choice – whether it’s an AI-enabled digital journey or an in-person conversation. We have committed to keeping our branch network open until at least 2030, alongside the adoption of new AI tools, in recognition that technology should enhance member interaction, not replace it.

3. AI hyper-personalisation

Third, AI enables building societies to move beyond one-size-fits-all products and create an even more personalised approach. For example, AI-powered financial coaching might nudge members about unnecessary spending or missed savings opportunities. Tailored affordability insights could help customers understand risk earlier. Or adaptive mortgage and savings journeys could adjust the level of support based on their unique history, circumstances, and preferences.

Members increasingly expect digital experiences to ‘know them,’ but the mutual advantage is the ability to do this responsibly, through transparent explanations, data ethics and human oversight. Personal AI should not focus on cross-selling; it should act as a member advocate, protecting their long-term interests and ensuring that humans can step in when empathy or complex judgement are required.

As personal AI becomes more embedded into everyday life for all generations, members may even start using their own digital assistants to manage money, compare products and interact with their building society.

The bottom line

Ultimately, technology may carry more of the operational load, but trust and human connection should remain the foundation of building societies’ approach to AI. To stay trusted and relevant, building societies must focus on using AI to add real value, making the technology clear and understandable, and enhancing the personalised experience that members already expect.

This article was first published by the Building Societies Association.