Machine learning in UK finance - hype or reality?

The purpose of Bank Overground is to share our internal analysis. Each bite-sized post summarises a piece of analysis that supported a policy or operational decision.
Published on 31 January 2020

UK financial firms are increasingly using machine learning to help run their businesses, and are exploring ever more sophisticated techniques. Such innovation can improve financial services, but authorities must remain on top of the risks.

Computer programmes can, with limited human intervention, recognise patterns from data and automatically make decisions. This is called machine learning (ML).

Two thirds of UK banks, insurers and other financial services firms that we surveyed are already using ML to help run their businesses (Chart A).

Chart A

Two thirds of UK financial services firms surveyed use machine learning applications

For many of these firms, ML projects have moved beyond the initial development phase. Our survey results show that, in the majority of cases, ML projects are live within the business. Firms are using ML in a range of tasks, such as credit scoring, securities trading and anti-money laundering checks.

Firms are also using a variety of ML methods, including complex ones, such as neural networks and natural language processing (Chart B). Their complexity means they can be more difficult to interpret than simpler techniques. Firms we surveyed use, on average, a combination of three methods per use case.

Chart B

UK financial services firms use a variety of machine learning methods (a)(b)

(a) Firms often use more than one method at a time which is why the percentages add to more than 100.
(b) The underlying data are based on the use cases provided by survey respondents.

ML has the potential to improve financial services, including better personalisation for customers and lower costs for businesses. At the same time, we need to consider the risks it can pose. This is why we, alongside the Financial Conduct Authority, are establishing a forum to engage with firms to support the safe deployment of ML across the UK financial sector, so that we can all benefit from its use.

This post has been prepared with the help of Carsten Jung and Oliver Thew.

This analysis was presented to the Supervision, Risk and Policy Committee in August 2019.

Share your thoughts with us at