An intelligent future for banking processes
Last year, Barclays announced that they envisage a future where robots will do our banking. An Artificial Intelligence (AI) system similar to Apple’s Siri could be used to converse with customers, make recommendations, and receive information.
‘We’re very soon going to be entering a world though, where we may not have to be physically touching a device in order to execute transactions or to be able to engage with computers.’ – Derek White, former chief design and digital officer at Barclays.
The idea that Artificial Intelligence and robotics could take over banking processes, eliminating the need for human employees is certainly not new, however, progression in technologies, combined with greater computing power and greater availability of data are allowing what was once an expensive and complex answer to both managing and improving banking processes to be a practical reality.
Until recently, businesses have largely used automation to remove the human aspect of processes, replacing simple human tasks with clear rule based decisions with computer logic. So what is the difference between simply automating processes, and combining with an AI technology?
Definition – Artificial Intelligence: the theory and development of computer systems able to perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.
The ability of Intelligent Automation (the combination of artificial intelligence and automation) to recognise patterns and adapt to new situations is the differentiator between simply automating a process to cut costs, and really giving banks the edge to optimise and redesign their processes. It’s the progression from process improvement to complete process overhaul. AI leverages new software capabilities such as language processing, speech recognition and, fundamentally, machine learning to achieve unprecedented levels of efficiency and quality in processes which previously would have been carried out by humans; what’s more – the robots can do it better.
There has never been a more relevant time for banks to embrace the use of AI. Traditional banks are under threat from fin-tech and challenger banks, which often differentiate themselves by utilising newer technology, and are under pressure from numerous regulatory demands that can often introduce new and complex processes across the bank.
In order to retain clients, banks are focusing on front end processes to improve and personalise the customer’s experience. Artificial Intelligence could be the key to transforming many of these crucial customer facing processes and retaining the competitive edge. For example, UBS uses AI to deliver personalised advice to the bank’s high net worth clients by building a profile of an individual based upon modelling millions of individuals’ behavioural patterns.
Regulatory processes can be effort intensive and rarely merge seamlessly with the bank’s day to day activities. AI could be embedded across these processes to ease the burden on individuals, whilst improving the processes themselves. For example, software such as Feedzai is now commonly used in banks to identify fraud and financial crime through predictive analytics that are far more effective than human data analysts could be.
Over time and as the technology becomes more powerful, we could see a scenario where an AI implemented process is able to monitor, manage and redesign itself as the demands of the environment in which it is placed change. But today, is it palatable, or even possible, for banks to relinquish control?
Firstly, AI is limited to specific portions of specific processes – it can’t yet be used for everything humans can do. Compared with automation, it is unlikely that AI would be able to replace an end to end bank process as it will usually only take over one ‘intelligent’ part of a process.
Secondly, to ensure regulatory compliance and retain overall control, the underlying processes would need to be mapped, analysed and constantly measured to confirm that the AI is performing as expected. Banks would need to focus on the development of management information dashboards, key performance indicators (KPIs) and key risk indicators (KRIs) to fully measure the benefits, and control any anomalies and new risks introduced by an AI process. Controls and reviews of the output would be of paramount importance to mitigate any risk of a process going down an unintended path, and to measure the benefits of AI to the overall process.
Finally, current processes would need to be adapted in a way such that the AI software would be able to slot in. To see the greatest benefits of the system, processes need to adapt to the AI technology, rather than the new system adapting to the old (possibly inefficient) way of working. It is unlikely that the surrounding processes would remain optimised with the addition of AI.
So what does this mean for us humans? Despite a reduced headcount, it is clear that humans will still be required for the foreseeable future in the underlying process management of any AI system. Through careful planning and thorough process analysis to control the integration of the technology, there is huge potential for banks to turn what may look like futuristic dreams of artificial intelligence into practical and cost effective realities.