sdecoret - stock.adobe.com
The robotic process automation market is shifting.
RPA software and services are expected to grow to more than $22 billion by 2025, according to a recent report by analyst firm Forrester.
However, the RPA software portion of the market will only be worth about $6.5 billion by 2025, while the services market will be valued at $16 billion, Forrester said.
Despite the overall growth in the RPA market -- largely fueled by pandemic-sparked demand for software automation -- the market is expected to level off due to the sharp rise of intelligent automation, said Forrester analyst Leslie Joseph, the report's lead author.
Intelligent automation is the combination of RPA and machine learning. The technology gives organizations the ability to have a machine learning-based capability that could, for example, scan an invoice, extract data from it and pass the data downstream to an RPA bot.
In this Q&A, Joseph expanded on the plateauing of the RPA market and what leading vendors have begun to do to prepare for it.
Why do you expect the RPA market to level off?
Leslie Joseph: RPA always was meant to be brittle technology. It was not meant to be durable in the sense that you create a native integration between two applications and enable a workflow that flows through the state of integration.
RPA has to evolve for two reasons. It's not technology that you would want to run your enterprise in the long term.
Leslie JosephAnalyst, Forrester
The second reason is that, if you look at organizations and the tech service team within which they've automated using RPA, a lot of these organizations have started in the back office -- finance, accounting, human resources procurement. They are rich and fertile grounds for RPA because they're consistent processes.
But then, when you're done automating all this low-hanging fruit and you're looking for the next set of processes to automate, you're typically going outside of these common, consistent processes into more complex processes, a lot of which are industry-specific. They require a lot more capabilities than just simple UI layer integration to create long-term durable solutions. And that's where RPA starts to break apart because it's not able to handle these kinds of situations.
As the market levels off, what will happen to vendors like Blue Prism, Automation Anywhere and UiPath?
Joseph: All these guys that are in the RPA business are critically aware that when automation starts moving to this more differentiation-centric, transformation-centric kind of construct, these guys are the first guys who are going to go under the ax. They are the most brittle form of automation. Nobody's building durable, long-lasting enterprise applications using RPA. They're just using it as a Band-Aid.
There's this whole movement within the RPA industry to start to integrate more capabilities to be an automation fabric platform more than just an RPA platform.
UiPath is a very interesting company because it's a leader in the RPA space, but it's also wrapped up in this whole transformation. Over the last two to two and a half years, they acquired a process mining company, they acquired an integration company to handle APIs -- they acquired ProcessGold.
That transformation outward from being just an RPA company to being an integrated automation company and now slowly trying to become more of an automation fabric platform, UiPath is a very good bellwether for that transformation.
As the RPA market levels off, will there be a need for more people with technical skills?
Joseph: It's not really accurate to say, throughout the evolution of RPA, that it was always business user-friendly and anybody could code. There was a certain amount of skill required; you needed to know a language -- C++ or what have you -- to be able to code a bot.
Now, that has grown simpler [and] the learning curve has grown shallower, primarily because RPA vendors have tried very hard to make their tools more accessible to business users.
That has come a long way since the beginning. So, even when you talk about AI, there are obviously data scientists who are building models, but then you also have consumers who don't really care about things like model drift or feature engineering or whatever you want to call it. Instead, all they want is here: 'I've got this workflow, and I want to put this little model in here and start to make these decisions.'
There is a distinct evolution of the kinds of personas that are creating applications today, and that evolution is felt not just within RPA, but across the entire spectrum of application development, including AI and machine learning.