RPA is too narrow for today’s enterprise

(Rob Enslin of UiPath during our video interview)

Do you remember when we debated the merits of UiPath as an RPA vendor – and watched the drama of a massively-hyped IPO unfold? That IPO was only two years ago, but in enterprise software years, it seems a lifetime.

Fast forward to today: UiPath now positions itself as an AI-savvy automation vendor: UiPath Reveals Expanded Generative AI and Specialized AI Offerings. I see only one mention of RPA in that news release, and that’s in the context of UiPath’s “AI-powered Business Automation Platform.”

Do workers welcome the automation economy?

Remember when automation debates centered on job loss impact? In May, UiPath issued fresh data on the “automation generation,” a report with an altogether different narrative: New Global Survey Reveals Nearly 60% of Workers Believe AI-powered Automation Improves Job Fulfillment.

Nearly 60% of respondents believe that automation can address burnout and improve job fulfillment, and 57% report that they view employers that use business automation to help support employees and modernize operations more favorably than those that do not.

Why the attitude shift? Because workers are being asked to do more with less – a lot more. Human or machine, they’ll take all the help they can get:

Employees are being asked to do more work with less support, with 28% of all global respondents saying they’ve been asked to take on more tasks at work in the past six months because of layoffs or hiring freezes. As work pile-up takes a toll on employees—more than one in four workers (29%) around the world report feelings of burnout—more staffers are leaning on AI tools to provide relief.

We could argue these attitudes must be revisited with generative AI in mind; I can’t think of any workers who savor the potential of AI taking over the vast majority of their jobs. But that’s a debate for another time (I don’t see that kind of job loss anytime soon). The point for UiPath is: the automation playbook has fundamentally changed.

UiPath – leadership changes pending, but how did we get here?

Well, with the possible exception of investor market drama. If you do a search on UiPath stock, you’ll see a confusing array of graphics and charts – some trending up, some down, depending on the outlook mood or criteria. Add in a pending change at the top: on July 11, 2023, co-founder Daniel Dines announced his intentions to (soon) step away as co-CEO, and into the role of Chief Innovation Officer: Founder’s Update: UiPath Announces CEO Transition:

With that, today we announced that – following a transition period through our fiscal year end of January 31, 2024 – Rob will become sole CEO and I will assume the newly-created role of Chief Innovation Officer.

So what’s next? No one better to address this than my next guest on the proverbial hot seat: UiPath co-CEO Rob Enslin, soon to be the sole CEO.

My history with Enslin goes back to spirited discussions when he was in leadership roles at SAP. But SAP has a way of sticking around, as per UiPath’s recently-deepened SAP partnership. So I asked Enslin: how does he assess his first year with UiPath? Enslin’s UiPath story goes back further, even before Google Cloud, when he was still at SAP – and more than a little bit skeptical of so-called “robotics process automation.” As Enslin told me:

I kept getting questions to go speak to Daniel Dines –  from everybody. This happened a couple of times while I was at Google Cloud as well.

I honestly did not want to get into RPA,” Enslin joked, chuckling at the unexpected turn of events. Why?

I knew RPA from my days at SAP.  I always felt SAP should have developed the majority of the functionality that RPA was in 2016/2017, which was invoice matching, kind of simple stuff. I viewed it a little bit as screen scraping, to be honest. And so I wasn’t that interested.

A closer evaluation while at Google Cloud elevated Enslin’s perception:

We did a due diligence, when I was at Google Cloud, with a couple of RPA vendors, and we actually found out that UiPath, technically, was superior to anyone else in the RPA space, which was quite interesting to me.

“That’s what excited me – to build out the automation category”

UiPath kept popping up. Enslin would hear about nifty customer use cases (“I started hearing a lot about UiPath… customers were very happy with it.”). Eventually, a UiPath board member convinced Enslin to spend time with Dines. That was when Enslin’s eyes were truly opened; an untapped opportunity loomed  large:

At that point in time, I realized UiPath had much more than a RPA product, that it was probably as close to an automation platform as I had seen… I knew there was a market for automation – a market that needed to be consolidated. There was a category called automation that was really fuzzy in the market, fuzzy for users and fuzzy for customers.

The analysts didn’t really help, because they never really had a clear definition of it either; there were five or six magic quadrants and things running around. So I said to Daniel Dines: ‘I will come along. But we have to go off to automation, not RPA.’ That’s what excited me: to build out the category of automation.

But hold up: what differentiates an automation product from an automation platform?

The reason I thought we had a platform? When I first came to UiPath, we had something called a tapestry, and I couldn’t understand it. I looked at all the solution sets. We had acquired process mining [ProcessGold]; we had task mining. My view is: every company needs to go through a discovery process to understand what processes they are doing; where are they ineffective? How do you actually make certain you know what to do with them – not only uncover opportunities in processes, but actually make the changes, so that businesses would benefit from it?

Enslin had firsthand BPM experience from the earlier BPM hype cycles. But this wave of process mining felt different:

I was fascinated by task mining, because I always felt that the one piece missing in the whole BPM modeling and process mining was the user – how they user behaved, and how the user interacted with systems, and: how do you actually make certain that it’s not just the systems, but it’s the actual interaction with the system that changed?

So I view that as unique. And then we acquired a company called Re:Infer in the communication mining space – think emails, chat – where we’re able to actually bring that into the whole process of discovery, and include that in the discovery phase of determining where to automate, where to improve processes, where to improve how manpower operates, and impact that. And I said to Daniel. ‘Are we onto something here’?

Enslin knows how big SIs close out enterprise deals, and where the sticking points are. He told Dines how this would go over:

I can tell you the big GSIs will love to work with the solution, because they will be able to really go down to a depth. At the same time, we have this incredible technology that you bolt, which has basically been known as RPA, but it’s got a low-code platform in there. It’s got a workflow platform in there. It’s got development environments for citizen developers, for analysts, for professionals, and so on. Today, we call that the Business Automation Platform. And on top of that, you have your testing environment.

My take – on market position and generative AI futures

UiPath can’t get comfortable now; the needle has moved again. We can’t have a conversation about these topics today without examining the automation possibilities invoked by generative AI. Suddenly, vendors have a different kind of AI scorecard, and no one knows exactly how to judge it (I’d hate to be the one tasked with maintaining quadrants on this right about now!).

Then there is the financial health of UiPath, and how Enslin assesses their overall position. Now that Enslin has been in the co-CEO seat for a year, how does he assess the financial market perception?

I’m probably always pretty critical on myself, no matter what. No matter what success looks like, it feels like it always can be better. When I came in, the pandemic stocks were taking a massive hit at that point in time… There was the Russian situation that just happened. That was actually evolving as I was joining UiPath. At the same time, we were talking about transitory inflation, or some kind of inflation.

Which became the “macro economic headwinds” we’ve heard reference to in countless software earnings calls. Enslin says that for UiPath, the currency issues were no joke:

We had a lot of currency headwinds; we moved out of Russia. We do business in local currency, which at times is good for you, and at times bad for you. It was always a significant headwind for UiPath, because the business was pretty large in Europe, and in Japan in particular. 

Investor criteria has shifted also; now there is more of a bottom line orientation:

UiPath, at that point, had not really turned a profit… Before I joined, everything was growth, growth, growth… We can’t spend everywhere; we’ve got to choose where we’re going to spend. You need to show investors you are innovating.

And how did that work out?

I feel really good that we actually now are relevant. Because our product should be relevant in these kinds of markets. Our solution set should help companies be more successful, be more efficient, utilize resources more efficiently and so on.

UiPath had to change how it communicated a value prop to executives, and change it yesterday:  

I kind of intuitively knew that we had to be in the C-level suite, and we needed to get there fast, as fast as possible. Otherwise, it would be really hard to work in this environment. If we’d started a year earlier, it would have been great, but I think it was timing was [still] great.

But investors turn on the markets, and the markets are impatient. Enslin explained that UiPath committed to turning a profit, and they did.

 We were actually a Rule of 40 company in Q1. We won’t consistently be a Rule of 40 company, but we committed that we will be Rule of 40 company.

Enslin shared more profitability stats; I plan to revisit that in a further piece, for now, the UiPath web site has updated financials from Q1. Investors seem to like the UiPath financial narrative, but from my own  subjective interpretation of buy/hold/sell ratings, investors’ growth and profit criteria aren’t satisfied yet. Then again, you can lose your customer focus if you obsess on short-term Wall Street expectations. For me, digging into customer project success is the most important barometer. Enslin loaded me down with field stories and results; I’ll digest those and issue an update.

As for generative AI, it’s too early to judge an enterprise software vendor on their generative AI functionality. We’ll get further into UiPath’s AI strategy in future pieces, but UiPath isn’t new to integrating machine learning into its workflows. On the generative AI front, UiPath’s expanded AI announcements from June 27 cites UiPath’s OpenAI connector, already in use by some UiPath customers.

Generative AI is sure to have an impact on automation – whether it’s 2x, 3x or 10x and beyond remains to be seen. One thing I know: the original RPA vendor approach was too narrow, and too brittle. But generative AI can be brittle too, when it is confronted with scenarios beyond its training data. Yes, reinforcement learning can mitigate that, but to what extent remains to be seen. In other cases, generative AI lacks the proper guardrails and enterprise data sets to live up to its hype. Few customers have the data science resources to navigate all that on their own; they are counting on vendors like UiPath to get that sorted.

Enslin is determined to make good on an automation platform that avoids the narrow/brittle pitfalls of RPA. I go back to the need for a deeper process analysis of what is sub-optimal, and what is a prime candidate for automation – that remains a constant. Show a customer how you can identify and solve that, and they won’t care if it’s AI, or RPA, or some other acronym not yet (ab)used.