Manish Sood, CEO at Reltio, explores how banking enterprises can get their data in order to harness and utilise AI in the most effective way
The global banking industry is finally catching on to Artificial Intelligence (AI), digital and cloud transformation, thanks partly to digital upstarts that have found innovative ways to grow amidst challenging markets. For example, digital-only and some data-driven regional banks have grown their deposits by as much as ten percent. In contrast, traditional banks have experienced deposit declines of three to five percent over the same period, according to research by McKinsey. Top-performing institutions have excelled by focusing on relationship banking and superior digital service in consumer finance while using data-driven interest rate pricing as another example.
The global banking and financial services sector has taken note of these opportunities. It is undergoing rapid digital transformation driven by the adoption of AI solutions. According to The Bank of England, 75% of financial organisations were using AI in 2024, with a further ten percent planning to use this technology over the next three years.
However, this surge towards using AI risks being undermined by a perennial problem in enterprise IT – a proliferation of siloed data linked to legacy and modern applications and systems. Whilst much of the impetus for AI in banking is to improve operations and customer service by collecting and analysing data, the reality is that for too many banks, their data sources are fragmented and incomplete.
So, how can banks ensure their journey to effective AI usage is not delayed or derailed?
The obstacles to effective AI-driven enterprise data integration
Firstly, business and technology leaders within banks must acknowledge the barriers to AI deployment. Whilst it is one of several issues, including low trust, a lack of explainability in the technology, and integration issues, siloed data that is poor quality and incomplete is the most fundamental problem to correct.
How data is siloed compounds the challenges of banks getting their AI-driven enterprise data right. A data silo is a collection of information that is stored within a specific application, department, or system that is not easily accessible or shared across an organisation. Data silos often emerge as financial services organisations add specialised tools to solve specific business challenges. Here, each tool generates and stores its data, leading to greater fragmentation and more data that is held in silos.
As well as this, as banks scale in size and add more applications to their back-end systems, it creates even more data stored in its hubs within the business. This continues to snowball as financial organisations rely on multiple third-party platforms, all of which create their own data and store it in their systems. There is a lack of inter-collaboration between all parties, the bank itself and external partners. Reltio’s 2024 Data Leader Survey found that 82% of respondents said that over 40% of their organisation’s data is derived from over 50 applications.
Consequently, data silos prevent banks from gaining a complete, and therefore accurate, view of their operations and their customers. As such, this makes it challenging for these financial services organisations to harness data for AI-driven insights, efficient decision-making and seamless customer experiences.
Overcoming the challenges with a unified data strategy
No matter how ambitious a bank’s AI plans are, they often find their organisations are stalled by inconsistent and fragmented data. So, whilst financial services organisations pursue AI-driven transformation efforts, many remain entangled in a web of disconnected data, jeopardising their success. Even those businesses with mature data governance frameworks face persistent data silos that hinder their AI initiatives.
So, as banks continue to spend their ever-tightening budgets on digital transformation, they must ensure that they are making the most of these efforts. The Global Banking Benchmark Study uncovered that 32% of executives said budget constraints are a significant barrier to digital transformation. So, it is important for banks to break down these data silos to improve their AI-driven initiatives. Not only that, but financial services organisations must also ensure that the data, which is fed into these models are high quality, complete and trusted.
Therefore, financial services organisations must have a strong data unification strategy, or they will be at risk of their data silos becoming that much more troublesome as they continue to scale and grow. It will also have a significant impact when it comes to providing service to customers. This is because data which is stored in silos will not give that customer service agent a full view of the customer and so will not be able to offer personalised or proactive service. As such, all data about each customer should be stored in one place which is easily accessible.
Banks should focus on driving customer-centric initiatives, with a strong focus on customer 360 views to enhance experiences and create operational efficiency. So, it is gratifying that most of the respondents in the aforementioned survey of data leaders across all sectors, including banking, have plans to upgrade data architecture within the next 12 months to improve the unification and management of data across their enterprise.
As financial services organisations continue to grow, scale and add more third-party applications, their siloed data issues will only continue if banks scatter valuable information across increasingly fragmented sources. With a modern data unification tool, banks can consolidate data and deliver trusted and timely insights to drive best-in-class customer service. As banks continue to further their AI-driven digital transformation efforts, only an integrated data unification solution will unlock AI’s full potential and empower banks to thrive.