Is RPA a stopgap in automating complex banking compliance processes?
Robotic Process Automation (RPA) is proven to deliver cost-savings whenever it is used to automate simple processes. However, for more complex banking processes it comes with a warning to avoid over-utilising it, and, the predominant view is that enterprise architects should focus on using robust, scalable, APIs connected using proven and secure technology. Yet RPA still has its supporters in the banking and financial services world – despite the complexities of Know Your Customer (KYC), Anti-Money Laundering (AML) and Enhanced Due Diligence (EDD) processes.
Accenture Consulting’s whitepaper, ‘Evolving the AML journey: Operational Transformation of Anti-Money Laundering Through Robotic Process Automation’, begins by highlighting the cost-benefits of RPA. It also claims that RPA can boost “throughput and improve quality, while remaining compliant across the AML ecosystem.” The introduction of the 2016 report cites other benefits too: they include an implementation period of just 6-8 weeks, higher staff satisfaction by “eliminating monotonous tasks and allowing individuals to focus on higher value work”, reduce incidents caused by human error and – perhaps controversially to some – fewer full-time employees will also work on repetitive tasks.
Nothing new about RPA
Francis Carden is the Vice President of Digital Automation and Robotics at Pega, which acquired robotic automation software provider OpenSpan in 2016. Carden, former Founder and General Manager of EMEA at OpenSpan, adds that there is nothing new about RPA, which used to be known as desktop automation. “It wasn’t called RPA until 2012 and in 2005, when we started OpenSpan, the technology was so different to what had gone on in the past.” His team was able to raise the necessary funds to kick-start the firm.
Moving on to the end of 2019, he finds that there are two key trends. The key one that RPA has, in his opinion, been over-hyped in terms of its capabilities to the extent that he says that some analyst reports predict RPA’s demise by 2021. He explains why:
“It is a recognition that there is a significant lack of scale. Few companies using traditional RPA have more than 50 bots, and so people are waking up to realise that it has been oversold. This is backed up by industry research. The plethora of applications of the desktop don’t sit about saying, ‘automate me please.’ The banking and financial services industry were the first to get into RPA around 2013, and now they are waking up to the fact that RPA is not the promised Holy Grail.”
Highlighting bad processes
Instead, the implementation of RPA is making organizations aware of bad processes. Yet with the drive to deliver efficiency and cost-savings, Carden says: “People are looking at processes and thinking how they can automate them to improve the customer experience. They thought that RPA would buy them time, allowing them to push back digitizing the processes. However, the banking and financial services industry needs to offer better customer service, and the biggest threat is from the newer entrants that enter their market that are born truly “digital.”
The future of RPA in KYC and AML
- RPA is useful for pulling together parts of an IT estate you don’t control – i.e. internal databases that sit outside of a bank’s core KYC or CLM solution
- RPA is not very effective without an orchestration platform – a solution that helps coordinate complex tasks which may vary, depending on process, risk level, entity type, jurisdiction, source type and user role
- For example, tracing ultimate beneficial ownership up a complex tree in multiple geographies – might require leveraging a combination of APIs, RPA and/or customer outreach
While RPA can work well with some simpler processes, Carden explains that it’s often just “too hard” when it comes to more complex ones. This is because a complex process may involve a large plethora of complex applications (one bank has over 6,000 legacy applications to manage), which have often been created and customized many times during the last 30 years. He therefore thinks of Robotic Process Automation as often simply the automation of computerisation, and he regards computerisation as what we used to call the automation of old manual (pre-computing) labour. Consumers today live in an “I want it now” culture, meaning that nobody really wants to think about waiting more than a few days for a mortgage application to be processed. Applying RPA to these older processes might get you there a little faster but often these processes are just not applicable any more for a digital world.
He adds: “The big banks are stuck with legacy systems. If you are a digital bank and you just built the systems, you’re not tied down by legacy systems. This doesn’t change the underlying cost. A digital company can therefore get stuff done, but we’re not going to get digital with RPA.”
This leaves the question about whether enterprise architects should develop new APIs and technologies for the efficient management of KYC, AML and EDD processes. While you can automate anything you want with RPA, Carden finds that long term, the costs and overheads can often outweigh doing it the right way (using APIs). “The reason they like RPA, but fear digital transformation, is because they think RPA is quick and has high value,” he says. They also think that digital transformation will take too long.
Carden elaborates: “KYC, AML and EDD are really for the new century. The old systems made it difficult to track your customer, and RPA isn’t the magic fix. People are waking up to realise that digital transformation technologies are becoming more cost effective and quicker than trying to just apply RPA to everything. If you have existing APIs, you would use those, but that doesn’t involve RPA. Developing new digital capabilities is quite revolutionary because it is much more agile than the way banks traditionally work. Having a strategy to close the gap is not a bad thing.”
Low Code and No Code
In Carden’s view there is currently a revolution happening in the way applications are built. This involves Low Code and No Code application development models. He says they permit enterprises to “rapidly build new digital processes faster than trying to wrap the old ones with RPA.” To him they equate an application development revolution. He argues that the financial services industry must recognise that there is a new way forward, given that “organisations want to upgrade and enhance their systems on-the-fly, while people are using the applications, like most digital companies are doing today. You can also build something that’s truly modular, unlike the monolithic applications of the past.”
Many modern KYC, AML and EDD solutions are built on No Code or Low Code platforms. He advises enterprises to buy technologies that are built on these platforms because they offer security and compliance out-of-the-box and are flexible (agile) to the extreme. He explains: “This allows you to bring together all of the benefits of low-code with AI and machine learning too, from a single platform. You don’t have to wait 6 months anymore; you can build new processes and applications in days, and you don’t have to write any code. Hard coding just isn’t necessary anymore and business leaders are reaping the benefits”.
RPA sticking plaster
So, will Robotic Process Automation ever be suitable for complex banking and financial services processes? Yes, but it shouldn’t be used as a plaster to cover poor processes. Carden likens this to people putting wallpaper over wallpaper until all the layers eventually completely fall off. In his opinion it’s time to “stop papering over the cracks.” However, he says that RPA can help with the digital journey and act as a backfill. Whilst this amounts to “turning robots into poor man’s APIs”, it can be a justifiable bridge.
Graham Jarvis is the Managing Editor of Media-Insert Communications and is a widely published business, marketing and technology journalist.