Generative AI (Gen AI) technologies are rapidly reshaping the financial services industry, transforming areas such as distribution, financial advice, and risk management. According to Maxim Afanasyev, Financial Services Industry Lead at Google Cloud, successful adoption of Gen AI typically involves addressing several pillars, including business value use cases, continuous data management, and AI governance.

Afanasyev categorises business value use cases into three primary areas:

Reimagining Distribution and Financial Advice:
Gen AI enables banks to move from a physical-first model to a digital-first approach, while still providing personalised customer experiences. This shift helps improve financial advice and guidance through more efficient digital channels.
Automating Middle and Back-Office Operations:
Gen AI can assist in re-engineering an institution’s cost structure, even in the face of legacy technology and siloed organisational structures. This transformation automates middle and back-office operations, enhancing efficiency and reducing operational costs.
Improving Risk Prediction and Data Management:
As financial institutions face increasing complexity and fragmented data, Gen AI can enhance the management of financial and risk data. This technology helps predict risks more accurately, improving decision-making in an environment of growing uncertainty.

However, continuous data and robust AI governance are crucial to ensure compliance. Regulators are playing an active role in supporting financial institutions, as seen with the Monetary Authority of Singapore, which launched Project Mindforge to explore the risks and opportunities of Gen AI in the financial sector.

Afanasyev notes that businesses have multiple options for accessing Gen AI models. Organisations need to carefully consider the pros and cons of each approach. For instance, when a Gen AI model is downloaded into an on-premise environment and trained on internal data, it may become obsolete over time if it does not update with new public information. This could result in business users complaining that the model’s outputs do not account for new regulations, news, or changes in the industry.

Afanasyev advises organisations to rely on Enterprise Gen AI models to ensure up-to-date information and continuous learning. Additionally, organisations must be cautious about where their data is processed. Data sovereignty regulations may impose restrictions not only on data storage but also on data processing outside specific geographical boundaries, which may affect where and how data is handled.

In conclusion, while Gen AI offers significant opportunities for banks to enhance their operations and customer experience, its successful implementation requires careful consideration of data governance, compliance, and the continual updating of models.