AI offers significant value for financial services with product maturity improving

Artificial intelligence (AI) is steadily revolutionising financial services. According to global accountancy body ACCA using advanced technology could transform financial services products and services in a way that has never been possible before.

Narayanan Vaidyanathan, Head of Policy Development, ACCA, said: 'A key differentiator and value proposition for AI is the ability for personalisation at scale. Historically, the choice has been high volume and standardised; or low volume and customised. AI allows for the ability to 'learn' about the characteristics of a customer over time and create solutions that will map specifically to their needs - and to do so at scale. It also allows for a range of use cases, with some examples particularly relevant for ACCA members including contract analysis, fraud detection, invoice coding and bookkeeping, forecasting/predictive analysis and narrative analysis of financial results.'

However, in response to a consultation by the European Commission's (EC) Directorate-General for Financial Stability, Financial Services and Capital Markets Union, ACCA did warn that care was needed over the quality and accuracy of the output of AI.

Fiona Murray, Head of EU Public Affairs, ACCA, said: 'Many AI solutions offer tremendously powerful value propositions. But the day-to-day performance and delivery doesn't always match the expectations (perhaps even hype). This is often due to gaps in the training data, data quality not being maintained over time, or the model not being tuned/retuned as often as it needs to be.

'However, this can be mitigated by effective management of surround factors such as the human-in-the-loop touchpoints and organisational processes around the technology itself.'

With the EC looking at how financial service sector businesses are developing or planning to develop or use AI applications, ACCA outlined three key advantages for businesses looking to invest in the technology:

• Greater alignment between solution and customer need based on learning more about customer profiles over time and reflecting those dynamically in product characteristics;
• Less demanding upfront capital expenditure requirement for many AI solutions, as licence/subscription models mature.
• Lower barrier to entry for those who are not expert programmers with the option of low code/no code platforms, and coding assistance from Generative AI.

Narayanan Vaidyanathan added: 'AI is the best technology we've seen so far for dealing with unstructured data and could help with areas like sustainability reporting - extracting information, analysis and structuring reporting templates for compliance. On financial data it is already used for maintaining/updating the general ledger, with possibly greater support in the future for areas like preparing statements and conversion of statements from one standard to another. However, it's important to stress that human judgement remains a key aspect and AI is still capable of errors linked to insufficient understanding of context.'