There are three dominant players in the burgeoning robotic process automation sector, which is increasingly providing potentially game-changing efficiency tools, mostly for large companies.
Blue Prism, the oldest and original market leader, coined the term “robotic automation” in 2012. Automation Anywhere emerged several years ago as its top competitor.
But the youngest member of the triumvirate, UiPath, has lately caught up with the others and is now the fastest-growing. In January, the company hired Marie Myers, one of its early enterprise users and a strong RPA advocate, to be its new CFO.
“I came to this job because I really love RPA,” says Myers says, who had been global controller at HP Inc. “I had other offers, but I turned them all down — I didn’t want to be the CFO of a hardware company or a plumbing company. Scaling for growth is a big challenge, but I wake up in the morning excited, even though I know it’s going to be a long, tough day.”
She views RPA as something that will improve the workplace and workers’ lives. Part of what’s on her mind with regard to that is her three kids.
“After I graduated with my third university degree, I got a job as an analyst, but I worked 14-hour days and did a lot of cutting-and-pasting,” she says. “Now, imagine a world in which I could have worked half the hours but done interesting and meaningful work, with all that manual stuff gone from my plate. That could be the life of my kids.”
During a recent visit at CFO’s New York office, Myers talked about dealing with hyper-growth, UiPath’s strategies, emerging trends in the RPA space, and her views on the robotics software market. An edited transcript of the conversation follows.
What’s required to scale the company for the kind of growth it’s envisioning?
The challenge is getting the team wired for what that requires and then building out the company’s operational underbelly. We’ve already scaled from basically zero; we hit $175 million in annual recurring revenue at the end of 2018 and we see a very fast growth trajectory for 2019 and 2020.
Before you came to UiPath, it was your job to deploy RPA at a company. Your new job probably requires you to have a bigger-picture view of the RPA space. Where are you on that curve?
Well along. I had done RFPs with all three major vendors. Right now, most companies are just scratching the surface on RPA. They’re in the proof-of-concept stage, with maybe three or five bots in the organization. Very few have gone to an enterprise scale with bots, where they’re either deploying hundreds of them to run parts of the company or providing a [dedicated] bot for each employee.
A bot for each employee? I’ve only thought of bots as automating various processes, not being deployed on a per-employee basis.
Think about what you do in your day. I’m sure there are manual tasks, whether it’s cutting and pasting data from a spreadsheet, or maybe cutting and pasting voice transcripts into documents as text. Or even just sending automated birthday cards or filling out your gym schedule.
There’s only a handful of companies in the world that are thinking very creatively around how RPA can be a key element of their digital transformation. A few are looking at enabling each employee in order to raise their level of digital literacy and understand how to use these digital tools.
I think that shows how broad the market can be. It started with desktop automation many years ago. That was the genesis of the [business process outsourcing] industry, where a lot of labor was moved to lower-cost markets. That has matured into RPA, which replaces a lot of that manual work performed by humans.
Could that particular creative use of bots, deploying them on a per-employee basis, take off and become a major trend?
Absolutely. Just the way touch-screen on cell phones did. Initially, bots weren’t seen as enablers. In fact, they were viewed somewhat negatively, as something that would take away people’s jobs.
It hasn’t been that way in, for example, Japan. People there are more comfortable with bots being part of workflows. There is a more mature perception developing that RPA can really change the way work is done.
I have a bot on my laptop. I hit the run button in the morning and it performs a number of functions. For example, it scans my email, runs some sensitivity analysis, and pops to the top the most important messages. It runs through expense reports that need to be approved. So it helps to prioritize the tasks of my day.
So there are various tasks that bots are programmed to perform automatically and others that you can instruct them to do at a particular point in time?
Exactly. That’s the difference between what’s called attended and unattended bots.
Back to the idea of RPA replacing jobs. We hear all the time about how it’s not about reducing headcount, but rather freeing up employees for more strategic, value-added roles. But are the accounts payable, accounts receivable, payroll, and procurement folks really going to be able to fill those roles?
In my background, almost everybody we hired was a university graduate. I myself once had three degrees but was building pivot tables as my entry point in the workforce. Did I want to do that? No, but I had to in order to have a job.
Is there a level of reskilling and retraining that’s required to make this successful? Yes. You can’t just plug and play. Someone who has been doing a routine job can’t just wake up the next morning enabled. But I think most folks in finance have a level of education where they can think about this as an opportunity.
Looking again at Japan, it’s a very mature country with a significant shortage of workers. They see RPA as a productivity lever. It’s not a unique case, but it’s a good one to study because it provides insight into the kind of transformation that can happen.
What do you think the relationship between RPA and AI will be? Will AI absorb RPA? Or is RPA destined to become more AI-like?
I see RPA as a gateway to enabling AI. In the beginning, RPA was kind of bespoke software on people’s desktops and laptops. In the past few years, it’s come along as an automated platform not only for the development of bots but also for the ability to govern and monitor [automation].
AI is kind of where desktop automation was a decade ago. The models again are bespoke, although they’re sitting in the cloud or on servers. You don’t necessarily have robustness in terms of inputs, governance, and controls. But RPA can be applied very scalably to managing AI. We’re working on exactly that product right now, where you’ll be able to bolt AI components onto RPA.
That’s important from a finance perspective. You don’t want to be relying on AI for elements of your financial processes, cybersecurity, or IT infrastructure without robust governance and controls.
You mentioned earlier that few companies right now view RPA as the linchpin to a broad digital transformation, but rather use it to automate isolated processes. What is the event or thought process that has to happen to trigger a move toward that broader usage?
Thinking about my own journey, it started when there was significant pressure on results. When there’s an operating cost issue on the table and the company needs to think through a massive transformation, some of the traditional tools won’t work anymore. What makes RPA attractive for that purpose is that it’s fairly easy to implement and adopt at a low cost.
Aside from scaling, are there any notable challenges or strategies that might be unique to UiPath or to the RPA space?
First of all, I’m dealing with a new category. How do you define it to analysts and investors? What do the KPIs of success look like? There is a set of KPIs for measuring the performance of SaaS companies, but while we kind of look and feel like one, we’re not really a SaaS company.
Then there’s building the financial acumen — the business telemetry that feels right for the category, without getting too much into predefining what we should do based on what everybody else thinks is right.
You can take a classic business model and apply it to this model, and you might get it wrong. That makes me stay up late at night and think. Try to build a predictive model when you have only two years’ worth of history. Is that enough to base a forecast on? How useful is the traditional finance mindset of a three-year plan?
Does that uncertainty require you to follow a number of different strategies at once, hoping one of them will prove to be right? Or do you need to settle on a single strategy and do everything you can to execute it?
I’m not a big fan of running multiple strategies. I think you have to focus on a core strategy, but you have to be agile enough to make changes frequently. And “frequently” may be every other month. For example, we pay out sales compensation once a quarter. But we might switch to once a month, which is pretty much unheard of.
We have to be agile because of the rate of technological change and the speed of this business. Traditional budgeting, planning, and forecasting processes don’t really work for hyper-growth.
Do you ever get to go home?
Not much right now. I’m hoping that gets better. But I love RPA so much that it’s not really work to me. At HP I had two and a half thousand employees, and RPA made their jobs a whole lot better.