From intelligence to ability: how Barclays is harnessing AI
The capabilities of AI tools are quickly evolving. But what will it mean for financial services? Craig Bright, Group Chief Information Officer, considers how applications of artificial intelligence are impacting the sector – from customer support to fraud protection – and explains how Barclays is using these innovations to strengthen its services.
It’s no secret that AI is transforming the banking world, bringing speed and depth to a myriad of processes. However, it’s how banks apply this new technology that will determine what this means for their customers.
Many years ago, when I was working for a power supply company in engineering, I first encountered AI – and it fascinated me. As software developers, the problem we were trying to solve was how to capture the expertise of specialists in power distribution to allow the team to deliver electricity more efficiently. We created an expert system that would help a less experienced operator optimise the performance of the furnace for generator power.
That was a way of making a system more efficient, so people could leverage their expertise and provide better solutions elsewhere. This, in a more sophisticated way, is what AI can do in financial services. The fundamental ambition of working with people and bringing a higher standard to the work they do hasn’t really changed. But the technology has become dramatically more sophisticated.
Barclays has been a technological innovator throughout its history, from developing the first British banking computer to opening the first ATM to creating the first British credit card and launching the first UK contactless payment card. Generative AI could be as transformative as any of these innovations – bridging the digital and human sides of our operations.
Fuelling AI through data
At the bank, analytical AI has been in use for some time. This ‘traditional’ AI focuses on performing a specific task intelligently, that is, making smart decisions according to a specific set of rules. For example, we’ve been using this type of AI to manage risk within our organisation, such as through voice recognition software to help enhance security for customers.
However, generative AI takes things a step further by learning the patterns of a store of data and generating something new. Chatbot services are probably the most well-known example of generative AI, but the technology’s applications in financial services present impressive opportunities as well as challenges. A 2023 report from McKinsey has found that generative AI could increase productivity by 3% to 5% – and reduce banking operating expenditures globally by up to $300bn.
To get the best out of generative AI, you need context, and context comes from data. This is where Barclays has an in-built advantage. We have the opportunity to leverage our store of data within AI and with over 48 million customers in over 40 countries – the sheer amount of information that we manage means the insights our AI tools draw on are current, relevant and bespoke.
In recognising that data – is the fuel for AI, we’re investing in modern cloud-based platforms, creating high-quality data products and boosting data literacy across the organisation to help create value from this technology.
Staying on the “winning side” of AI innovation
While we’re looking forward to making the most of the opportunities that AI offers, we’re also mitigating the threats that can accompany them.
All new advances can be used by ‘bad actors’ practising phishing, creating deep fakes or trying to undermine biometrics. It’s a constant battle that we know we have to be on the winning side of to help protect our customers.
While analytical AI has long helped us manage fraud and money laundering risks, generative AI creates opportunities to help us reduce these even more – something that is absolutely essential for us as a bank.
If we think about fraud as a set of rules, AI allows us to read patterns of behaviour that produce warnings. Fraud alerts typically come from detecting unusual behaviours. If the signs from the data show that the behaviour isn’t unusual for me, I’m less likely to get an alert, while if an action is seen as ‘normal’ for most people but strange for me, it can be detected. Generative AI means that risk assessments for the customer are being brought to an individual and safer level.
AI doesn’t mean letting go of the human side of our services – instead, we aim to give our customers the convenience of a digital experience with the comfort of a personal touch.
Group Chief Information Officer, Barclays
However, as with any new technology, we have to test and adopt it responsibly in a way that aligns with our values – with security and customer experience at the heart of any decision. This means ensuring compliance with existing and emerging laws and regulations.
Of course, it’s important to note that AI doesn’t mean letting go of the human side of our services – instead, we aim to give our customers the convenience of a digital experience with the comfort of a personal touch. The opportunity lies in how we use AI alongside everything we know about our customers to improve their experience with us. While each successive generation of AI can more effectively use natural language and provide quicker responses, it isn’t intended to be a substitute for human contact in those moments that matter.
On that point, AI can make our digital support capabilities more intuitively intelligent and enable them to answer customers’ questions. But it can also provide our colleagues with the information they need to respond to people efficiently, so their focus can be on establishing connection and empathy with our customers rather than spending time searching for information.
If our customer service agents are asked a difficult or specific question, the AI acts as an assistant and boosts the human agent’s knowledge. The AI is working alongside them – and that has enormous potential because banking is a complex industry where responding quickly, clearly and personally to complicated questions can sometimes be difficult.
The more we can do to help our colleagues, the more seamless, relevant and timely our customers’ experiences will be. In that way, AI can increasingly allow our customer service teams to free up their time, serve more people and get more done.
That applies across the bank, where we are working with our major partners to test out AI-enabled efficiencies. With Microsoft Copilot, we are exploring the potential to drive productivity across our workforce – using its generative AI capabilities to summarise meeting actions and synthesise detailed reports, for example – while AWS’s Bedrock helps us improve frontline customer engagement. And it’s not just about supporting frontline staff – we have 30,000 technologists, so improving their efficiency is a huge enabler for the organisation.
By reading data well enough to understand our customers’ contexts, behaviours and relationships with us as a bank, we can use AI to help make sure Barclays is providing the best possible personalised service.