Companies are deploying robotic process automation (RPA) software at a significant pace, with year-over-year enterprise spending on RPA increasing by double digits. IT research and advisory firm Gartner projected that the global RPA market will reach nearly $1.9 billion in 2021, a 19.5% increase from 2020. Gartner also predicted double-digit growth rates in the market through 2024.
However, there's a shift in how that money is being spent. After years of organizations deploying RPA on premises or within their own private clouds, they're moving to SaaS RPA.
In its "Emerging Technologies: RPA Software Advancements" report, Gartner predicted that more than 20% of RPA deployments will be cloud-based by 2024, up from 1% in 2020.
Figures from RPA provider Automation Anywhere Inc. are even more dramatic. The company reported in a January 2021 study that 67% of new customers opted for cloud deployment, and in August projected a 72% decrease in on-premises RPA deployments over the upcoming year, making on-premises deployments just 10% of the total.
"Vendors are maturing their cloud platforms, and we'll see more adoption of cloud-native RPA platforms as organizations recognize their flexibility, but that's three to five years away," said Ranyah Salous, intelligent automation practice lead for consulting firm Guidehouse.
The late adoption of SaaS RPA
SaaS RPA is a late entry in the enterprise race to the cloud, since on-premises RPA software was once the only solution and has remained the dominant option in recent years. Arthur Villa, an analyst with Gartner, explained that much of the reasoning has to do with how RPA works and what it does.
RPA creates software robots, or bots, that mimic human interactions with software. These bots basically perform high-volume, repetitive tasks that don't require any decision-making or complex thought. Simply put, RPA creates automation at the desktop level.
"Execution of those automated actions happens on premises nearly all the time," Villa explained.
Ranyah SalousIntelligent automation practice lead, Guidehouse
Moreover, many organizations initially used RPA for quick digital wins, and some still do. They often utilized the software to work with legacy systems that didn't integrate easily with cloud solutions. As such, many enterprise leaders were satisfied with on-premises RPA solutions -- even as SaaS options hit the market.
"Companies with on-prem systems and on-prem RPA [deployments] would probably need some overarching reason to move to the cloud," said Eric Dynowski, managing partner and chief solutions officer with Deft, an IT advisory and service provider.
Additionally, experts said enterprise leaders often wanted on-premises RPA because they believed it was needed to keep latency to a minimum and maintain the desired level of security.
"Many [organizations] wanted to retain control with their own security policies," Salous said. They might host RPA solutions on their private clouds, but it has been rare for them to adopt a SaaS RPA. "When hosting on their own cloud, they got to maintain control of the configuration and controls and didn't want to lose that."
She estimated about 80% of her clients have RPA running either on premises or in their private clouds.
"Very, very few clients actually use cloud-native RPA platforms," she added.
The outlook for SaaS RPA
That dynamic, however, is quickly changing as more RPA vendors offer products fully hosted and managed in the cloud.
As Villa noted, RPA vendors brought their offerings to market after seeing the expansion of cloud computing and the growing demand for SaaS products.
Today, major RPA vendors -- including Automation Anywhere, Blue Prism and UiPath -- offer RPA as a service. Additionally, cloud providers have introduced similar capabilities with offerings such as Microsoft Power Platform to enable users to automate processes.
Analysts and IT advisors said the benefits and drawbacks of hosted RPA solutions mirror those typically listed for other SaaS offerings. Fully hosted and managed RPA products require no installation and maintenance work from the enterprise, letting organizations move more quickly than they could with on-premises products.
Although some organizations remain hesitant with RPA as a service because they think there could be some latency issues or security drawbacks, those concerns have not played out, experts said. Costs were also cited as an issue: Cloud RPA can be more cost-effective than on-premises RPA but not necessarily when used in large-scale deployments.
Evan Campbell, managing director at Protiviti and part of the consulting company's digital practice, said organizations with existing on-premises deployments may not want to automatically shift to RPA as a solution. He also noted some organizations may have RPA working with legacy systems that may not integrate easily with newer SaaS options, giving them reason to stick with their on-premises deployments.
However, in line with the aforementioned research, Campbell predicted RPA as a service will soon be the dominant choice for new deployments, noting that his own company is adopting RPA in the cloud.
He said he doesn't see latency as an issue for RPA, as typical deployments don't need near-instantaneous responses, nor does he see security issues with cloud-native RPA platforms. Still, he does see upsides to RPA in the cloud, including quicker startups, an easier ability to scale and limited maintenance requirements for IT staff.