Artificial Intelligence, Robotic Process Automation, Intelligent Automation, Blockchain, Tokenisation, Machine Learning - banks are struggling to work out which technologies are most suitable for their needs and which they should prioritise in order to get
the most out of their investment. In some ways, this has always been the case with technology, but there are some differences today.
Focusing on automation in general, research by
Capgemini estimated that the financial services industry could expect to add up to $512bn to global revenues by 2020 through ‘intelligent automation', and there is no doubt there is potentially a lot to be gained by its implementation when done well.
Robotic Process Automation (RPA) - usually referring to point automation solutions covering one area or sub-process rather than an entire journey – has delivered improvements but also customer and employee frustration.
RPA can assist customer service agents by, for example, automatically logging in to applications and ordering them in a way that helps agents navigate efficiently when serving customers. Removing these rather simple but frustrating sources of friction can
increase speed-to-outcome by allowing employees to focus on the customers rather than the applications.
What’s more, RPA can also offer a foundation for improving key customer service metrics. Using RPA and desktop analytics in conjunction with CRM, BPM, and service desk applications can also facilitate streamlined visibility and accountability to outcomes.
However, this is where automation is best - where it is built into the end to end solution rather than layered on top and has some intelligence built in. As banks focus more on effective automation, a blend of intelligence built into the automation, ensuring
that when needed you get help from the best skilled person, will become ever more important. We are all becoming more accustomed to things working really well in an automated manner, and any element of advice or exceptions management needs to be managed really
well to differentiate banks. This will be the customer service battleground. Errors and delays will become less and less tolerated.
Taking journeys or customer experiences in the context of end to end and focusing on exception-based manual processing augmented by technology will be the ultimate focus. High value activities, advisory, error or dispute processing, customer onboarding,
KYC and compliance activities in general will benefit most.
There’s a distinct split between scenarios with a clear processing flow that are driven by business rules and do not change frequently, compared with complex scenarios and where customers want and need to talk to someone live. In the most basic case, the
employee or technology receives inputs, examines those inputs, applies a rule to them, with no discretion in this scenario, and then sends the output forward to the next step in the process. But this covers activities that should be simple and, based on risk
appetite, be fully automated with some risk-based testing. This is happening already and being executed well at scale. Where the challenges now lie are to complete this across many areas, but more so to get the interaction right between automation and person
to person interaction.
Clients I deal with are already using Intelligent Automation (IA) to, for example, manage large volume channels such as email and picking out key words through natural language processing (NLP) to automate routing of an email or automating an action with
no intervention. This can often be decided based on whether there could be another opportunity to talk to the customer or not. For this, AI and machine learning can be applied to optimise the next best actions. And clearly, this is best done across all channels
to give a channel-less feel. There is huge potential across not only mass retail, but also wealth and commercial segments to better automate sales and service, yet not lose the human touch where it is warranted for the customer, the bank or both.
Customer friction is starting to disappear for account opening, savings and deposits, as well as payments reconciliation, fraud error and some simple lending. It still has a way to go for all lending, investment and advisory activities, but in many cases
some friction or intervention should be welcomed by customers when dealt with by the right skilled people in a timely manner.
I believe we will stop talking about different subsets of automation and refer to it more often – in the interim at least – as Intelligent Automation (IA), combining the best of analytics, AI, automation and people. Eventually, it all comes down to how well
automation is implemented in an end to end customer and employee journey. And maybe we will just call it that in the end.