By Kshitij Jain, Head of Analytics for the UK and Europe at EXL
Since the release of Chat GPT earlier this year, generative AI has moved up the boardroom agenda and adoption has accelerated across industries. Uptake is predictably higher in unregulated industries, where use cases are already abundant, but in financial services, the regulation landscape calls for a more considered approach.
An AI boom in financial services?
In fact, when compared to retail – a sector in which generative AI is already “pervasive” according to EY’s recent consumer index – adoption in financial services is perceived as lagging behind. However, despite the constraints that regulation may place on financial services institutions, there is an encouraging uptick in interest, and investment in generative AI technologies. This shouldn’t be a surprise. According to McKinsey,[1] corporate and investment banks (CIB) in particular first adopted AI and machine learning decades ago, well before other industries caught on. For example, trading teams have used natural-language processing (NLP) to read tens of thousands of pages of unstructured data in securities filings and corporate actions to figure out where a company might be headed for a number of years.
Certainly, the largest players in the market understand the potential of these tools and are already making moves to get ahead. Goldman Sachs has started using generative AI to classify and categorize millions of documents, including legal contracts, and PwC recently announced a partnership with an Open AI backed start-up to enhance its legal work and develop use cases for tax. It is clear that the initial scepticism we saw from regulated industries has now given way to guarded optimism, which is encouraging. But what are the core ingredients for successful, safe implementation which delivers tangible business and customer benefits?
Creating a strategic AI roadmap
Financial services institutions must start by defining how they want to integrate generative AI into their processes, and the scale of transformation they want it to bring about, with appropriate guardrails deployed.
Generative AI has the potential to transform operations in a multitude of ways. It can completely revolutionise processes, improve decision making, and transform how insights are generated. It can also empower customer care agents with a more complete and intuitive picture of the customer, thanks to its ability to summarise and categorise structured and unstructured data. This makes hyper personalisation much more tangible and supports better management of complaints. It is also a powerful tool for tackling fraud, cross-selling, debt collection, and acquisition – but attempting to deploy it for all of these things at once can be a costly mistake and could see banks and providers fall on the wrong side of the regulator. A phased approach, which allows for auditing, evaluation and a culture of continuous improvement where any mistakes or bias generated by AI are considered and addressed, is needed to set financial services institutions up for success.
Data Pragmatism
It is widely understood that Gen AI requires large volumes of high-quality data, which is verified, compliant with organisation wide standards and where the provenance is understood. This is simply not possible without making an honest and pragmatic assessment of the organisation’s data operations.
Financial services institutions must ask themselves; Is there a robust data governance framework in place? Does it accurately reflect the way the entire business uses and interacts with data? Effective data governance empowers an organization to trust the integrity of their AI and machine learning models by ensuring that their data originates from reliable sources. It also creates rigour and ensures that the models used are aligned with the organisations’ principles.
There must also be careful consideration about the limitations of the data the organisation holds. Often there will be gaps, or areas for improvement, for example the data may not be fully representative, there may be some challenges with uniting and contextualising data from across myriad of structured and unstructured sources. Understanding and addressing this before is even considered, is crucial.
Organisation wide buy in
The transformative potential of Gen AI has captured the attention of C level executives across all industries. They know that this technology has the ability to disrupt their business operationally, strategically, and culturally, but their perspectives on its application and the associated investment can be wildly different. Getting endorsement, support and understanding of the impact of Gen AI adoption at C level can be a huge challenge, particularly in highly regulated industries where compliance is a prevailing focus. However, provided there is a keen eye on the objectives, and impact on people, processes and the bottom line, a consensus is possible.
The next challenge is to ensure widespread understanding of the changes and improvements Gen AI will enable, for all employees. Much has been debated around the ability of AI to ‘steal’ jobs or replace people, and this has no doubt prompted caution and distrust. Yet the arguments ‘for’ its implementation are so powerful – the potential of Gen AI to fight fraud, more accurately assess risks, personalise customer communications and equip employees with analysed, contextualised information to enable more informed decision making, is unprecedented. If this is understood across the organisation, positive and compliant interaction with the technology will follow – leading to a far greater return on investment and significant business impact.
[1] https://www.mckinsey.com/industries/financial-services/our-insights/been-there-doing-that-how-corporate-and-investment-banks-are-tackling-gen-ai
