Blog article
See all stories »

What can we learn from AI and ML use cases?

According to a recent survey by the Bank of England, the use of ML technologies in UK financial services firms continues to increase: over 70% of firms which responded were using or developing machine learning (ML) applications, with firms expecting the number of ML applications to more than triple over the next three years. The reported benefits of ML technologies are enhanced data and analytics capabilities, increased operational efficiency, and improved detection of fraud and money laundering (Bank of England, 2022).

If you are in the lucky 70% or so of firms who have already implemented ML, you know that you are on to a good thing. However, it might feel like you have already applied ML to all the obvious use cases within your business. On the other hand, if you have not begun to develop or deploy ML applications yet in your firm, then it might all seem like a huge uphill struggle to even begin to consider it. Indeed, it would seem reasonable to imagine that the actual percentage of firms yet to embark on their ML journey is even greater than 30%, since these figures are based on organisations who responded to a survey about ML (i.e. demonstrating self-selection bias).

When considering new opportunities for ML – or more broadly AI – applications, whether this is for the first time or not, it is useful to consider how other organisations have successfully applied these technologies. Often, this information can be difficult to access, due to it being commercially sensitive. In cases when it is available, it can be buried in the body of reports, survey results or other documentation. The purpose of my recent review and appearance this month in London alongside Google, is to help others to overcome this challenge and to share a systematic understanding of AI and ML use cases in the financial services domain after surveying the literature.

I will present the synthesised summary which is grouped under three main categories: risk management, organisational / operational, and enhancing customer experience and engagement. As is the case with any literature review, decisions had to be made about the grouping, categorisation and inclusion of use cases and their sources. For example, for a broader review which also covers AI and ML algorithms and risks relating to the use of these technologies, I would recommend the recent report by the Turing Institute (Maple, et al. 2023).

The financial services sector

According to recent surveys, organisations within the financial services sector are increasingly adopting – and benefitting from – ML and AI technologies. However, one of the hurdles to AI adoption is the identification of appropriate use cases. In this article we have explored a range of use cases which can be broadly grouped into ‘Risk management’, ‘Organisational / operational’, and ‘Enhancing customer experience and engagement’. In some instances, it might be more useful to abstract away from specific use cases in order to use a more inductive approach. To help with this, I presented three broad characteristics of AI/ML use cases, namely ‘Business processes’, ’Data’, and ‘Task type’, along with corresponding examples.

A summary of ML and AI technologies and applications would not be complete without touching on the potential opportunities offered by generative AI. Although these approaches have existed for several years, it was late 2022 and the public beta release of ChatGPT by OpenAI and similar tools by competitors such as PaLM-2; that drew them to the attention of the general public and business leaders. Currently, such generative AI approaches have yet to feature in systematic reviews of AI and ML applications in financial services (although Buckmann, Haldane and Hüser, 2021 did review and identify limitations of the earlier OpenAI large language model GPT-3). However, in the interests of completeness, you need to consider some typical areas where generative AI technologies such as ChatGPT could be effectively applied.

I look forward to sharing detailed reviews soon including doing so at our Google event in London this month!



Comments: (0)

Alastair Gill

Alastair Gill

Principal Data Scientist


Member since

04 Mar



Blog posts


This post is from a series of posts in the group:

Artificial Intelligence and Financial Services

Artificial Intelligence and Financial Services

See all

Now hiring