Are Legacy Systems Hindering Banks' AI Transformation?

The banking sector has entered a new phase of Artificial Intelligence (AI) maturity, with over half (51%) firms planning to increase their investment, according to the latest Lloyds’ Financial Institutions Sentiment Survey. AI has become an invaluable technology for the sector, powering internal operations, servicing as well as customer-facing applications.

That said, to unlock the full potential of AI, banks need to evaluate their legacy applications and modernise any that may hold back their transformation plans.

The AI Journey so Far

In line with customers becoming more aware of the potential benefits and more comfortable with what AI has to offer, everyone’s expectations for a seamless digital banking experience have grown. They expect apps to anticipate their needs while virtual assistants solve their issues in real-time. Banks are listening to these requests and are increasing their AI investment to further enhance customer interfaces and support them in managing their finances. After all, satisfied customers are more likely to trust and stay loyal to their bank.

Along with customer experience, internal operations and risk management are some other key areas that AI has heavily supported in the banking world. For example, loan and mortgage processing has been streamlined with intelligent document processing and machine learning minimising the need for manual data entry and credit assessment. The technology is able to quickly evaluate credit quality and provide insights in real-time, as well as automatically populating onboarding forms and KYC (Know Your Customer) workflows. However, there are still examples across all banking segments of retail, commercial and wealth where opportunities still exist to eliminate unnecessary tasks, improve quality and guide and coach employees and customers through processes. There’s still a huge efficiency, quality and revenue impacting change that is possible

When it comes to risk management, AI has been adopted to boost banks’ cybersecurity and fraud detection strategies, and to improve the accuracy and service levels of dealing with issues fast and accurately. Using AI algorithms, firms are now able to help shield their employees and customers from cybersecurity threats in real-time and create the necessary tools to help them avoid suspicious activities. AI systems are also able to identify and prevent financial crimes such as money laundering and impersonations. We are already in a world where we have AI tools checking and looking for AI generated threats. Hey, even my nephew has his coursework checked that AI has not generated it!

AI tools have so far made a huge impact in some areas on how banks operate, both internally and externally, but it’s still early days and legacy systems remain an issue. Legacy platforms can slow down this transformation journey due to their limitations in compatibility, efficiency, scalability and security. There are still many cases where you just can’t take advantage of the latest and greatest when you are held back by a system that is both old and not well understood. Modernising those systems should therefore be top of mind for any bank looking to introduce or increase its AI applications and usage.

The Challenges of Legacy Systems…easier than before

Legacy systems are deeply rooted within a bank’s architecture and have numerous dependencies that prove risky and complex to untangle. They usually have little to no documentation and are highly customised. Any reliance on outdated systems can cause unpredictable consequences to even the smallest changes.

In addition to this, maintaining outdated systems can be financially draining and time-consuming for banking employees, not to mention that remaining compliant is harder and can increase the risk of penalties and breaches. As a result, legacy systems can leave banks exposed to significant security risks and regulatory issues. However, the position we find ourselves in now is very different to even a year or two ago. There is far more capability to assess, analyse and then re-design more easily than before

There’s always an element of evaluation needed to identify which ones are important to keep, with systems often containing valuable data or supporting key operations, and which ones can be eliminated or updated. But once this is done, the ability to advance an AI transformation journey is now easier. To achieve this, banks can leverage GenAI and automation tools for speedy documentation and transition of their legacy systems, and then use design time GenAI to assess automation options and workflow redesign. This can genuinely be done in hours and days compared to weeks and months before.

It’s very interesting contrasting different bank AI plans and approaches. All have similarities at the core though and that is a desire to get really beneficial outcomes for customers, staff and the bottom line. Also an element of not wanting to miss out on the opportunity or be left behind. Some already have very good AI governance, programs and resourcing to succeed here. The technology that is chosen is important and a key enabler, but the change management and having clarity of purpose is still critical to success.