Cognitive computing evokes images of science fiction and endless possibilities that are in the distant future. However, a recent survey of banking executives suggests that 89% of those familiar with the concept thought that it would be very disruptive to the banking industry, and 79% thought it would be critical to the future of the industry. So what are the possible applications for cognitive computing in the banking sector?
Product and service development
Cognitive computing has the potential to manage large numbers of banking transactions in a much more sophisticated way, allowing banks to fulfil the promise of competing on analytics. It can use context- and evidence-based information to provide tailored products and services to customers. This will improve self-service for bank customers through enhanced automation of banking and investment products.
Investment advice has traditionally been the province of older, more affluent customers, often through independent financial advisers. In other words, it has been a fairly niche business, with a relatively small market. But cognitive computing offers the potential to provide automated but personalised investment advice on a smaller scale to almost anyone, opening up investments to a much wider audience. Having a financial adviser, albeit a robotic one, could soon be routine.
Cognitive computing can help banks and other financial services companies get closer to their customers. The ability to analyse text and other word-based inputs, and to learn from experience, adds another dimension for banks. Cognitive computing means the ability to improve personalisation with each transaction or interaction with a customer. And this, in turn, leads to huge improvements in customer satisfaction because banks are addressing their needs directly and accurately.
Cognitive computing can examine significantly more data, from more sources, including text-based and other non-numerical data. Because it looks wider than traditional analytics, it enables banks to get a better picture of potential risks, and therefore predict risk much more accurately. This improvement in credit risk assessment is a key part of digitalisation of the financial sector.
Previous fraud detection systems were often post-hoc, relying on analysing what had happened, and deciding whether it looked wrong. Real-time systems brought this up to date, allowing faster detection. Cognitive computing, however, could be used to predict fraud before it has happened, by looking at behaviour and detecting when it is atypical even before any fraudulent activity occurs. It could also have better potential for adapting to new frauds.
Management of fraud and potential fraud will also change. Cognitive computing systems can actually have a reasonable dialogue with people.This means that the computer could take its own action to manage a potential fraud, once detected. Instead of having to alert a human to act, it could contact the person concerned, and find out what has happened.
The use of cognitive computing frees up human fraud investigators for more complex cases. Since cognitive computing can detect and prevent quite a lot of low-level fraud and potential fraud, these cases will no longer need human investigation. This means that fraud investigators will have more time to investigate and manage more complex fraud issues—and we can be certain that these will follow.
Plus ca change
There are also downsides: cognitive computing adds its own security challenges. The day of the zombie baby monitors showed us how IoT could take down the Internet. Cognitive computing adds further cyber-security challenges as these systems can also be hacked. This may be particularly a problem in financial services if we are using these systems to detect fraud, because there is an obvious reason to target them.
It is also likely to lead to job changes in the financial sector. It is not just fraud investigators whose work will change as a result of cognitive computing. With computers taking over many of the transactions with customers, there are implications for call centre and other banking staff. The question is whether they will simply be redundant, or whether there will be a place for human interactions at the heart of banking and the financial sector. It seems likely that customers will want these human interactions, and that cognitive computing therefore represents an opportunity for these staff, rather than a threat.