UiPath sat down with Chris Lamberton, EY’s London-based EMEIA RPA Leader for Financial Services.
His role includes helping clients use Robotic Process Automation (RPA) to transform the agility and financial performance of their organizations, with a particular focus on bringing digital and cognitive capabilities to the customer relationship.
In this interview, Chris gives readers the benefit of his and EY’s extensive client experience as he expresses views on: how Robotic Process Automation enables insurers to overcome legacy IT challenges; what product innovations led to the current surge in RPA adoption; why many implementations fail.
UiPath: You’ve written about challenges faced by insurers in developing channels for effective customer communication — questioning whether traditional approaches can achieve the required transformation. Does that transformation still need to take place? If so, what role do you see RPA playing?
Chris Lamberton: Yes, it does, primarily because the insurance industry continues to be hobbled by legacy environments.
That said, these challenges apply across banking and capital markets; wealth and asset management, and many other industries.
As a leader in RPA implementation within the insurance and wider financial services industry, our clients include some of the largest general, specialty, and life and pension insurers — so I certainly see the issue first hand. Our insurance clients are being squeezed harder and harder by two pressing concerns: a need to innovate and a need to reduce costs. In both instances, they’re hindered by a multitude of 20- to 30-year-old legacy systems.
RPA is already playing a huge role in addressing those issues. This technology doesn’t just significantly reduce costs, but it also allows insurers to significantly improve services and deliver entirely new services.
As an example of RPA enabling new services, one particular challenge that is facing the insurance industry is that customers expect a good digital experience. They don’t want to be forced to talk to someone in a call center. Yet across the insurance industry, we think less than 10%–20% of products are digitized.
Some insurers can count over a hundred legacy claims systems, all of which are tied to numerous other systems. Of course not all the claims systems are 20 or 30 years old, but still, there is a lot of legacy. Finding the money to replace these legacy systems is hard, and providing a realistic business case for digital services in light of this is nearly impossible.
So, what’s the answer? We are now delivering projects using RPA technology to create simple ways to access and “wrap around” these legacy systems; then connect them to robot-aware digital technology to provide claims, and policy-change type services. So with RPA enabling digital, straight-through processes it becomes a tool for transformation and bigger financial and service benefits — bringing claims online; maybe providing customers with real end-to-end status on their claims, changing their policies and more.
Not only does customer self-service stop initial calls from coming into a claim contact center, but with a robot updating the web or mobile app with the claim status, customers calling in to chase claims reduces greatly. In our experience, upwards of 60% of calls in the claims area deal with one concern — “what’s happening with my claim?” so that’s a lot of calls that can be avoided. So, using RPA, EY is now enabling give new channels, much better customer service and significantly lower costs — that’s why we think RPA can be truly transformational.
UiPath: As EY uses RPA to wrap digital around legacy systems, is it seeing client benefits beyond end-to-end claim transparency?
Chris Lamberton: Absolutely, EY recently did an RPA project for a global carrier that involved wrapping the technology around their legacy environment. We took existing processes — where the claim settlement cycle was averaging 10–15 days — and showed how we could bring it down to 1–2 days. At the same time, processing costs were reduced by up to 90%.
Another key benefit is flexibility, because a robotic workforce can quickly adapt to change. Let’s say I change a process in a shared service with 100 people — this means I have to retrain 100 people and take them off productive work. If we estimate 80 people get the new process immediately, then 20 are still struggling for days or weeks. That, I would argue, is more difficult than teaching robots. While it may take a few days to teach a robot the new process, once one robot is trained, then every robot in the estate will do each task perfectly, every time.
Perhaps people are initially attracted to RPA by the cost savings. But while cost reduction is important, the technology can also make a dramatic difference in both quality and service performance, but also in its ability to offer entirely new services that just weren’t possible before. So we advise clients to create a portfolio of RPA initiatives that’s a balance of cost reduction, service improvement and service transformation
UiPath: What would you say is the ‘why’ and ‘when’ of RPA technology changing from a niche, swivel-chair automation solution to a significant force for labor arbitrage disruption?
Chris Lamberton: The financial services industry is certainly on an RPA journey. Some organizations are forging ahead on this journey while others haven’t started yet.
In my experience, early adopters view RPA as a real opportunity to gain important benefits; ranging anywhere from process performance to improved margins or overall competitiveness. Late or non- adopters, in my opinion, being held back by an “it’s too good to be true” mindset.
Of course you can’t blame them for reacting like that when you think about what we’re telling them — this new technology, RPA — will reduce cycle times, maybe pull a digital experience out of your legacy systems without a big IT project and, before we forget, slash your operating costs. It does sound a bit too good to be true.
RPA has a body of evidence showing it’s not too good to be true; it can do what it says it can. As the reality of its potential sinks in, more and more of our client engagements are not focused on convincing them it’s the real thing, they’re directed at showing clients how to get past the proof of concept stage and begin scaling real solutions.
Robotics truly enables the business as a whole, not just IT. In fact, one of the questions starting to emerge from business users is, “should we own our automation solutions?”. A business that can operate a virtual workforce of potentially hundreds of robots, delivering processes across the whole organization, may well have a big impact on their BPO service providers. Even today, many clients see RPA as a viable alternative to BPO, and that trend will only continue.
UiPath: Your partner role encompasses EY RPA centers of excellence. From that perspective, to what extent do you believe product innovation triggered or accompanied the move of RPA from niche technology to a mainstream automation solution?
Chris Lamberton: At EY, we view RPA technologies in terms of multiple generations and what we’ve seen so far are the first two generations
Generation one is what we call “Attended Robotics”, and it’s what most people think of when they hear the term, “swivel-chair automation.”
In this generation, the robots really worked alongside a person as a direct helper. Workers weren’t replaced, instead they were made more efficient, which limited scalability and cost savings.
There’s no question the people using robots were much more productive; but you still had people who can still be expensive. Attended RPA was also complex to implement and had a change impact on staff, so it has a smaller business case than the next generation — unattended robotics.
Generation Two is “Unattended Robotics”, where robots moved beyond being helpers and can execute process automation without the need for human involvement in low-value processes, allowing people to focus on higher value work.
From a technology perspective, providers began to introduce software and platforms that could scale to work across an enterprise, creating stable, adaptable “virtual workforces”. This is important because our clients are global corporations with stringent requirements for both operational benefits, but also security, audit, stability and scalability.
So, today enterprise-scale virtual workforce solutions are possible delivering significant cost reductions across workforces measuring into the hundreds of robots. And with a robot typically being the equivalent to 3-5 FTE or more, the scale of these robot workforces can make a huge impact on even the biggest organizations.
EY has seen real growth in this area. A year ago, there was very little choice of RPA products with these capabilities, but now there are more.
UiPath: Shared Service Centers (SSC’s) have generally incorporated BPO service providers into their business model and value proposition. Do you see RPA changing this paradigm? If so, how would you characterize these changes?
Chris Lamberton: Shared Service Centers were one of the earliest adopters of RPA technology; their motivation was margin increase and well as improved internal value from higher services levels.
The adoption of this automation solution by SSCs certainly puts BPOs in a difficult position. My position is that if you can describe a process well enough to outsource it, then you could well look at automating it.
We think that many BPOs believe their business model is under significant threat from RPA. That’s not to say some processes won’t always be more feasible to outsource than automate, and certainly the BPO industry isn’t going away anytime soon. But it’s hard for our financial service customers to look at their SSCs and think, “If we can automate these processes as well as our BPO, why do we need to have a contract with them?”
I know of some customers who have told their BPOs, “We know you can automate the processes we’ve outsourced to you. So we want to renegotiate the current contract down by 25%.” And in those cases, I think it’s possible the renegotiated contact won’t be renewed at the same scale next time around.
Now, I’m aware some BPOs are saying automation is actually a great opportunity for them to increase margins on current contracts — and it may be in the short term. And, some BPO’s are investing in their own robotic solutions but it’s hard to see these circumstances playing out well for them in the long term. This is tough for these companies - to have a business model that has constantly grown for the past twenty years, only to see it hit like this is hard. But that doesn’t change the fact that SSCs now have a choice — they can either develop their own virtual workforce or continue down the purely BPO path.
One significant factor in this decision is that SSCs can be much more efficient with their own robot utilization than BPOs can be with the robots used to deliver services to the SSC. When a SSC puts a process into the scope of a business process outsourcer, they can’t re-use any of those robots for processes outside of the contract. On the other hand, when all business processes are in the SSC’s scope, then every robot can be re-used for any processing across the entire environment.
This is a key distinction. I know of SSCs that share their virtual workforce across business units or even different operational companies. What’s important is those units may not have had any prior business reason to collaborate with each other, but now they’re able to take advantage of their collective RPA capacity and really optimize it.
EY can speak to this subject from our own operational experience. We’ve put robotics at scale into our own global SSC and learned it’s more than just taking people out of processes to achieve cost reductions. It’s also an opportunity to create a balanced portfolio of automation that produces savings, achieves service and quality benefits — then adds truly transformative services we couldn’t have done before.
UiPath: On the basis of EY’s work helping clients implement this technology, what would you say are the most prevalent RPA challenges and opportunities your clients either underestimate or overlook entirely?
Chris Lamberton: As one of the most experienced and largest RPA delivery consultancies, we’ve studied the lessons from a large number of engagements.
What we’ve found is that 30%–50% of initial RPA projects fail. Now, it’s not because the technology’s no good — after all, at least 50% succeed. And we’ve made RPA products work successfully across a landscape of legacy systems interlaced with every type of system imaginable. So it’s definitely not the technology.
The reason for this high failure rate is how it’s implemented. In November EY published a thought leadership article on top ten issues, and I’d like to share a few of those insights. We see a number of common mistakes. Some examples are:
Companies target the wrong processes. Often, they pick the most complex ones. Why? Because often those processes also cause the most pain. In reality, there is almost always much better business cases to be had with much simpler processes — especially when you’re just getting comfortable with the technology.
Looking at RPA only as an IT tool. It’s not. This technology can be delivered by the business with appropriate IT support and governance, but doesn’t necessarily need IT skills or IT delivery capacity. Also often we find that when RPA is viewed as an IT tool, the implementation reflects IT methodologies — which may be overkill for a tool that doesn’t change existing systems. The practical result is that automating processes that should take 2–4 weeks then take double or triple the time to implement using an inappropriate IT software methodology.
It can’t do everything - and shouldn’t try! Humans are still quite good at doing things that robots struggle with, so robots should be targeted at the 70% of a process that’s simple, with the 30% that’s complex left to humans. Trying to solve the complex 30% can take way more than 30% of the budget, and may still not deliver a good result – and you can always go back and automate more as you get more mature.
Also some companies have unrealistic expectations and find RPA can’t do everything, but may need to be augmented with additional tools, e.g., digital for straight-through processing, OCR for getting stuff off paper etc. So finding and using the right combination of tools for the job, and automating the optimum 70% of a process is key.
UiPath: Last question. When clients ask for a look into your RPA crystal ball, what does it show them for 2020?
Chris Lamberton: Busy! I’d say things might be going very quickly now, something like a rocket really — but they’re only just getting started. By 2020, we’ll have moved well past the question, “Does it work?” and onto, “Have I done enough?” The presence of robots throughout an enterprise will simply be a given.
Earlier I spoke of robotic process automation in terms of generations, with the first being attended automation and the second generation being unattended automation. The second generation — unattended automation — has had a tremendous impact on RPA growth, for the reason it has brought a 2-3x return-on-investment increase to automation solutions.
The third generation, which we’ve already discussed, is self-service. We think self-service will bring a further 2x to 3x Return-On-Investment increase to process automation, and we believe EY is one of the only organizations deploying these solutions.
The fourth generation will be cognitive technology. I fully expect it to expand the enterprise automation with capabilities such as AI-enabled machine learning and natural language processing. This cognitive generation could well make highly sophisticated rules engines a practical reality, allowing a virtual workforce to do more than ever. However, this cognitive generation raises a key question: how can this technology be utilized by business in a consumable way?
What I mean is, how will business use it? Right now, the first two generations have intuitive ways for business users to transfer business knowledge into robotic processes, without needing deep IT skills. But the cognitive generation’s technology could well be less user-friendly — and that could slow down its growth. The challenge will be for RPA providers to develop ways of interacting with these tools that allow business users to use familiar approaches; for example, the English language, to configure and revise complex rules engines, or simple machine learning interfaces.
UiPath: This scenario of sophisticated rules engines seems be far from the current RPA focus on mundane, repetitive, rules-driven work.
Chris Lamberton: It is, and it touches on a topic that warrants a discussion of its own; the question of how this cognitive technology will impact humans. You’re right; at the moment, people are concerned about RPA displacing jobs in lower-skilled areas of work, although there’s as many people who believe RPA liberates people to do the higher-value work they should be doing, not the manual work they currently do. But what about when automation impacts higher skilled work such as financial analysts or claims adjusters?
There’s no question robotic process automation will increase service levels and save money in these areas. However, it also means the person doing the manual work for the adjuster is gone. Now what if doing that manual work helped junior people start on the path to higher-paid, more senior claims adjuster roles? What will happen to skilled labor once the starter work is gone? How will people acquire the experience and judgment to replace the surviving skilled workers when they retire? That is a great topic in itself – maybe for another article.
UiPath: Chris, thank you for clearing your schedule for this interview. You’ve given me and the reader a valuable opportunity to learn from your RPA thought leadership and EY’s deep implementation experience. It’s very much appreciated.
Note: the views reflected in this article are the views of the author and do not necessarily reflect the views of the global EY organization or its member firms.
David Eddy has fifteen years of global application services experience.