Estimated reading time: 1 minute
Corporate banking is undergoing a transformative shift, with artificial intelligence (AI) and the array of next-generation technologies linked to it playing a pivotal role. However, the industry is still exploring ways to fully harness the technology’s potential; to evolve quickly with change while managing associated risks.
At a recent Finastra event, industry experts debated this topic. Below are some insights shared about the opportunities for AI in corporate banking and how to use it to unlock the full power of data.
Driving back-office efficiencies
When it comes to driving efficiencies in the back-office, Marc Smith, CEO of Conpend said, “The focus is increasingly shifting from robotic process automation (RPA) towards the ways in which AI can help with decision making. This is fundamentally transforming document processing and enabling straight-through processing. For example, imagine a world in which you submit a loan application, it’s approved, and you receive the money – all within five minutes. This is being made possible with AI. In fact, everything that is workflow-driven can potentially be replaced with a large language model.”
Dermot Canavan, COO, Trade and Trade Commodity Finance at ING, explains how AI is driving efficiencies in ING’s compliance checking processes. Canavan said, “Compliance checking has traditionally been an extremely paper- and people-heavy, tedious process, with lots of room for error. AI and large language models enable a much more automated compliance checking process, resulting in better data with fewer mistakes and gaps.”
As described in these examples, consider that corporate banking is very much driven by document processing. And to implement AI properly, complementary technologies are needed to make the overall strategy work.
For document processing, AI must be coupled with OCR (optical character recognition) technology to translate documents into data, and this will continue until such a time when the industry data standards can fill in the gaps.
Enhancing customer experiences
AI, and generative AI (Gen AI) specifically, could fundamentally change how corporate banks engage with clients. Enrico Muti, partner at McKinsey & Company, stated, “Gen AI can semi-automate pitch and proposal preparation, mine data to suggest client-specific actions, summarise insights from client interactions, and draft client emails. These technologies are also being used to vastly improve the ways in which chatbots can support clients with queries, acting as virtual assistants to help them navigate the bank offerings.”
McKinsey & Company’s partner Markus Röhrig added, “By automatically interpreting and completing technical documentation, Gen AI offers many opportunities to improve key processes like time consuming onboarding.”
Software developers are also using AI to enhance operational efficiency and elevate client experiences, such as writing code, creating unit tests, and automatically documenting work.
Implementing and scaling AI effectively: The three P’s
Russ Rawlings, Regional VP, Financial Services and Public Sector Leader UK&I, Databricks argues that to execute a data and AI strategy effectively, there are three key focus areas: people, process, and platform.
1. People
It’s important to consider the entire organisation when designing and implementing an AI strategy, Rawlings said, “This requires having a clear understanding of the roles and needs of all the different users of AI – including data scientists, data analysts and data engineers, as well as business users. You also need key people at every level to help drive change and ensure that AI is being embraced across the business.”
Although AI can automate manual processes and assist with decision-making, it still needs people to be truly effective. Canavan said, “The human element is always going to be key. For example, an AI system can extract data and correct it where it made mistakes, but you still need human oversight to ensure that process is robust – and to be able to step in at the appropriate points if it’s not going so well.”
Rawlings added, “To get everyone on board with the shift to AI, it’s imperative to communicate your plans, seek alignment on methodology, and build a strong community and culture where best practice and success stories are recognised and celebrated. Most importantly, it’s about being agile and flexible so that everyone can respond to changes effectively.”
2. Process
Adopting AI and large language models requires a robust plan with clearly defined processes. Not least of which coming from the fact that AI requires big data management and large computational capabilities, and this means that cloud adoption is necessary to integrate AI solutions. Separately, Rawlings said, “Implementing an AI and data strategy involves significant change, and planning the course of your change is essential. Big process change starts with defining the principles and rules; going back to those helps to make decisions and to move things along.”
Lewis Liu, Chief AI Officer at Sirion, added “This should include a framework to test AI models and clear rules to support validation. Human oversight and intervention are also crucial, particularly in terms of facilitating end-to-end data processing.”
Effective governance is also key. “To develop a clear governance framework for AI – and its associated data – an FI needs to understand what it’s doing, record what it’s done and have the evidence to support its actions. Good governance is about more than just ensuring you meet the necessary data privacy rules – it’s also about responsible AI, having high quality data, and ensuring the ethical use of AI,” said Canavan.
3. Platform
Finding the right platform is also essential for a good data and AI strategy. Businesses across all industries are increasingly moving towards platforms based on ‘data lakehouse’ architecture – an open, unified foundation for both unstructured and structured data.
The concept blends the benefits of data lakes and data warehouses, allowing institutions to better scale AI. Rawlings explained, “When built on an open foundation, a data lakehouse can easily integrate with the entire AI ecosystem to enable machine learning, business intelligence, and predictive analytics. What’s more, when combined with generative AI, this model provides a powerful data intelligence platform that can democratise data and AI across the entire organisation.”
Embracing the future
AI – and Gen AI in particular – is referred to by many as the fourth industrial revolution. Rawlings added, “It’s democratising data for users across the world. FIs can’t fall behind and must think carefully about how they embrace the revolution.”
Smith said, “Because of the amount of data they hold and the processes they manage for the economy, banks have a massive opportunity to move the needle. The future of our industry could look very different. So even though implementing change at this scale isn’t easy, the best thing to do is just to start – just do it.”
The key point here is moving forward, in small steps wherever possible. Legacy systems are a showstopper for AI adoption (and, in truth, most technology adoption), so there is a need for every stakeholder to be committed to staying at the cutting edge and to drive that culture.