Cracking the Code on How to Connect Modern Digital Platforms with Legacy ERP
Bank of Ireland shifts from human to robotic "workforce" for faster, better service
Although the results of SSON’s 2016 survey showed very little real RPA implementation, and perhaps even less understanding of its value, the discussion has gained momentum over the past six months. Indeed, at our North American Shared Services and Outsourcing Week the buzz was about nothing else.
The story in Europe is similar, and we look forward to engaging many of you in discussions in Dublin next month (Shared Services and Outsourcing Week Europe). In the meantime, we've spoken with a couple of people who have already implemented robotics and the feedback is, unsurprisingly, very positive. We’ll be sharing their experiences online in the next few weeks. Here, Brian Halpin, Head of Robotic Process Automation at the Bank of Ireland, talks about a pilot project turned success story with SSON’s Barbara Hodge.
One of the questions always raised around robotic process automation is just how ‘new’ automation really is. Many dismiss the novelty of the "robotics" moniker; others remind you that automation has been around since the early days of IBM. But there is a difference with RPA, and that difference becomes apparent whenever you speak with someone who has really ‘done it’.
Brian Halpin, Head of Robotic Process Automation at the Bank Of Ireland, makes the best case yet for robotic process automation. Where it really makes a key difference, he explains, is in its ability to connect a modern, web-based platform with a clunky legacy ERP. And as more and more businesses give in to the pressure to digitalize, and embrace mobile, virtual, and the cloud, it's shone a spotlight on the fact that traditional ERP just can't keep up with the demands of a digital workplace.
It’s a problem many businesses face and it can be resolved either by lengthy and costly IT integration projects; extensive and labor-intensive manual interventions (which has driven many to opt for offshore support); or, now, by a combination of robotics and minimal FTEs. The first option is generally too expensive and lengthy to consider; the second is one that many have chosen; but the third may well be the way of the future.
In the case of Bank of Ireland, the challenge was precisely to marry a state-of-the-art, digital, customer facing platform with the [limitations of] existing legacy systems on the backend. "Once we started hearing about robotic process automation we thought it was worth exploring as a possible solution to our problem," says Brian.
His team ran a pilot on a customer facing process around withdrawing money from a deposit account. This pilot proved enormously successful, not just from a cost perspective [reduction in required FTEs] but also by improving customer satisfaction significantly, and thereby reducing the number of calls coming into the contact center.
"The unexpected side-effect of RPA was that it effectively created a better customer service," says Brian, "enabling us to respond within 15 minutes as opposed to the 3 days it took in the past."
Where [RPA] really makes a key difference is in its ability to connect a modern, web-based platform with a clunky legacy ERP.
From the customer perspective, the service was suddenly that of "VIP" Brian explains. While the old manual process involved sending information and requests to the offshore center, going through quality assurance tests, and then sending the information back for customer communication, the reaction was now nearly instant.
"At the same time, we were able to release four of the five FTEs that were previously involved in this work into more value-adding roles," says Brian. "The proof of concept was very positive and return on investment took about three months. It was all incredibly quick and immediate."
So successful was the pilot, in fact, that Bank of Ireland is now considering RPA within some of its internal processes, too. Brian’s team is currently evaluating Finance and Risk as possible areas for RPA deployment. "It doesn’t really matter whether a process is externally facing or internally facing; as long as inputs are digital, and there is some rules based activity involved, a robot can replace a human more efficiently," says Brian.
A case in point is the possible support of risk management through RPA. Robotic automation would effectively gather data, reaching into different systems to look for aberrations or divergence from the norm, and monitor these readings. "In this case we would simply be reporting, not actually actioning a process as an output," explains Brian. "But monitoring risk robotically has a lot of potential."
The one nut to crack is to figure out where and how IT fits into RPA implementations. It’s somewhat of a grey area, as Brian readily admits. But while most organizations, including Bank of Ireland, have opted to place RPA into the operations themselves, IT is still needed for security clearance, ongoing support, and governance, although it's not necessary to run complex software development change models or to configure new processes.
One tip Brian offers to simplify any complexity is to maintain a consistent framework for the virtual workforce. "We didn't change or even improve processes as we brought in RPA," he explains. "Even if a new way of doing things might have been an improvement, we were sensitive to the fact that we wanted to use robotics to replicate exactly what humans had done." He says he constantly reminded everyone that ‘it's still a workforce’ and the same processes were being followed. This also satisfied IT’s security concerns as permissions around user interface did not change.
"If you remember that robotics simply does what people would otherwise do, you stay away from the potential minefield of thinking of RPA in terms of a technology or an IT project, which has an end date. The processes that are being handled require the same diligence and governance as before. The same owners are still accountable for performance. It’s just done with fewer people."
3 Tips: Things to Consider
- Break projects down into three-month windows, for easier assessment and control. Even if a job is actually much larger you can still fragment it into different stages to keep within that three-month timeframe.
- Don't aim for 100% automation up front. Go for, say, 60-80%, with 20-40% still being channeled through to humans. Then as things progress you can gradually move to a higher rate.
- Don't confuse RPA with "another piece of IT". The normal thinking is that technology is managed by an IT production team and Business Operations often relinquish responsibility for any breakdowns. RPA is really about a virtual workforce so it's important to keep the original owners involved to manage and be accountable for the process. Always remember that a robot is effectively a human in virtual form and requires the same monitoring and controls.