Artificial intelligence (AI) provides banks with the means to make better use of the massive amount of data at their disposal. In doing so, they can identify opportunities for growth faster and gain significant competitive advantage at a time when more agile digital counterparts are emerging.
I believe that without this insight, it is difficult to modernise processes or to understand how best to introduce new technologies into back-end systems.
When it comes to product development, the offerings of banks typically evolve slowly over time. Part of the problem is that legacy mainframe systems are expensive and slow to update whilst people resources for these systems are scarce (and becoming more so).
This creates a bottleneck in the process as the mainframe systems are typically the central record for all data. This means they must form part of any new solution developed. Front-end applications are relatively quick to develop but back-end integration or changes to core systems become the stumbling block.
Banks must therefore work with experienced and trusted partners to extract data from mainframes into modern database architectures that can use the data more effectively. In most instances, curated data extracts taken from the mainframe are very large, making it difficult for individuals to work with.
These are typically put into a spreadsheet format for users to work with. However, this is inefficient, prone to error and carries significant risk due to human involvement.
Instead, banks need the ability to extract their data and present this to users in a modern application interface for task processing. Risks are managed better, and operational efficiencies are dramatically improved when simple rule-based actions and decisions are removed from the human function.
People can then be empowered to add higher levels of value into processes through the analysis of data insights gained, a better customer-centric service model, and re-imagining of traditional processes. AI-led technologies enable this transformation.
Of course, all of this must be done within a strong governance framework. It is about building robust solutions that involve risk officers from early in the process to ensure all the necessary requirements are taken care of.
Getting insights from data is the first step to understanding where efficiency can be built into the internal processes of banks.
In many instances, banks have been relying on the same data processes put in place decades ago. These have gradually been tweaked over time without major overhaul.
Adoption of new technologies tends to be slow resulting in tool-sets, like traditional spreadsheets, remaining the primary environment used to analyse data-sets. However, this is neither efficient nor secure.
Even so, one of the most significant challenges revolves around exception handling. Most banking transactions require little human intervention. However, in cases where items are flagged (for example, AML/Fraud/Screening checks), this requires human intervention. Given the high volumes involved this comes at significant cost to the organisation.
It is therefore an excellent place to deploy new AI technologies such as Intelligent Process Automation (IPA). This streamlines processes and automates steps usually performed by people. Think of AI virtualising the human experience.
It is about building technology solutions that consider the exception handling process and automate as much of this as possible, introducing efficiencies and mitigating risk simultaneously.
Furthermore, in many instances, banks require multiple levels of human approval for transactions to be cleared. This is another time-consuming process that can be transformed through AI.
Instead of having two or three people view each transaction, the AI process can deliver the same level of expertise consistently in real time to improve SLA management and improve service to customers.
It is not about reinventing the wheel but optimising the robust processes banks already have in place when it comes to managing their data and processes.
Through refining and automation of processes a significant amount of human activity can be taken out of the system and this can be re-purposed to give the bank more capacity to focus on areas such as improved customer service or the design of new offerings.
Using modern applications that can integrate with existing data and processes, banks are able to generate insights from start to finish.
For example, look at the typical ATM infrastructure that must be managed daily. Transactions and GL account balances must be reconciled to ensure machines are working correctly, that no fraud is taking place, and there is always the right amount of cash available for banking customers without over exposure of capital reserves.
Using people to reconcile and investigate discrepancies is slow and inefficient. But using AI toolsets mean these tasks can be managed consistently, at high speed, and with full auditability.
Volume or capacity constraints are then no longer an issue. This extends into customer service as well, improving the customer experience when queries or complaints arise as there can be immediate action taken rather than waiting for a human to perform analysis and then take a decision.
An AI layer can be implemented to sit on top of existing processes while integrating into back-end legacy systems to deliver the value banks require.
Banks have high quality data, but it is not always accessible. Using AI to help manage the high data volumes can bring about significant improvements in operational efficiency which will ultimately deliver a better customer experience.