Unleashing the power of banks’ data with generative AI

The implications of generative AI on business and society are widely documented, but the banking sector faces a set of unique opportunities and challenges when it comes to adoption.

Avanade’s latest research found that bankers view automation and efficiency as the biggest benefits of generative AI – with AI having the potential to fundamentally change customer onboarding and fraud detection,  as well as automate regulation and compliance requests.

But despite this desire to unleash the full potential of AI, almost half (49%) said they did not fully understand generative AI and its governance needs.

Potential pitfalls

Generative AI also has the potential to cause issues if not implemented correctly. Unwanted problems include ‘hallucinations’ – incorrect information presented as fact – or ‘black box syndrome’, where it is unclear how AI decisions have been made. If banks are to put their faith in AI, then transparency will be key to building trust.

 It’s all about the data

 In order to unleash the power of your data, it needs to be AI-ready. That means moving your data into a dynamic cloud environment as well as consolidating data from the multiple sources where it is kept around the business.  If the data is not accurate, or comes from disparate sources which present conflicting pictures, the resulting decisions could be financially and reputationally disastrous.

This is a problem banking leaders are increasingly aware of. The report found almost two-thirds (63%) do not completely trust the data their company uses. It also found that just over a quarter (27%) completely trust their ecosystem partners to fully protect their customer data.

Banks are particularly vulnerable to this problem because almost 90% of the top 100 banks still use mainframes to run complex workloads related to their core business processes. These systems tend to be highly customised, complex and expensive to run, with data essentially ‘locked away’ where it is difficult to share. At best, data is held in ‘islands’ and silos, some of it unsecure, some in legacy systems and some of it not fit for purpose.

Unleashing data with generative AI

Fixing this problem is paramount. Banks must ensure they prioritise what data they need, before ensuring it is AI-ready. Doing so can have spectacular results.

A European bank wanted to reduce churn in its mortgage business, increase fraud detection and improve customer lifetime value. We built a customer data and analytics platform. 100+ variables were fed into machine learning models that predicted the risk of mortgage churn and fraudulent credit card transactions. Mortgage churn was reduced by almost 50% over a 6-month period. The client saw a 2% reduction in underwriting fees and a 7% increase in the early detection of credit card fraud.

 The potential of generative AI in banking is immense.  Unlocking the power of your data is the key to success. For more information read our report; ‘Banks: are you AI-ready?’