How Gen AI is transforming banks’ financial advice, risk, and operations

Gen AI can be used to reimagine distribution, improve financial advise, and transform risk management.

Generative AI (gen AI) technologies are transforming the financial services industry, and successful adoption journeys tend to cover multiple pillars and include business value use cases, continuous data, and AI governance and business process transformations.

Business value use cases can be structured into three groups, said Maxim Afanasyev, Financial Services Industry Lead, JAPAC, Strategic Industries, Google Cloud. First is in reimagining distribution, as well as financial advice and guidance. For example, this may involve moving from physical first to digital first whilst still delivering a personalised customer experience.

Second is using gen AI to automate middle and back-office operations. This involves re-engineering cost structure while saddled with legacy tech and siloed organisation structures.

Third is improving risk prediction and transforming management of financial and risk data. In this case, gen AI can be used to manage risk effectively whilst facing growing complexity, fragmented data, and manual processes and controls.

Continuous data and AI governance is critical to ensure that financial institutions are compliant, with regulators taking an active role in supporting financial instructions, Afanasyev said.

The Monetary Authority of Singapore, for instance, has launched Project Mindforge, which looks into the risks and opportunities of gen AI for the financial sector.

There are multiple options to access gen AI models, and organisations should understand pros and cons, Afanasyev said.
For instance, a GenAI model downloaded into an on-premise environment to be trained on internal data might not update with public knowledge following the download and become obsolete.

“There might be complaints from the business users that the outputs of those models might not capture new public information such as new regulations or news,” Afanasyev said.

Organisations should remember to rely on Enterprise gen AI models, said Afanasyev. 

Afanasyev also told organisations to look carefully at where their data is processed. For instance, data sovereignty regulations might restrict not only storage of some data but also processing of data outside of sovereign boundaries. Or restrict data from being processed in locations which do not meet certain criteria.