The Automation We Need Is Not The Automation We Deserve
At its TechEd 2018 Barcelona event last month, SAP announced that it is adding robotic process automation (RPA) to its software suite, starting in early 2019. Two aspects of this announcement are particularly interesting.
The first is that SAP has chosen to build its own RPA capability instead of partnering with any of the existing product vendors. The product demoed at TechEd contains most of the functional elements of a standalone RPA platform: a design environment; a central repository to store and manage bots; a “marketplace” for third-party bots by partners; and a runtime environment, supported by process monitoring, analytics, and audit trails. The product also integrates with SAP’s evolving machine-learning platform and toolkit.
The second aspect that’s interesting is how SAP has chosen to implement its vision of automation. Unlike typical RPA products in the market, SAP is focusing on native API-based integration. This also implies that, at least to begin with, SAP’s version of RPA will be limited to SAP’s walled garden only, beginning with S/4HANA and then later on extending to SuccessFactors, Concur, etc. It will forgo screen-based integration with third-party and legacy IT or business applications. SAP execs we spoke with talked down the current types of “unintelligent” screen-scraping-based RPA software in today’s market. However, they did concede that extending to non-SAP applications is something that the company is looking into; they just don’t have a public road map yet, and it may not be soon.
The RPA Of The Future Is A Feature, Not A Product
Robotic process automation is on fire. Just the three Leaders that we identified in “The Forrester Wave™: Robotic Process Automation, Q2 2018” have raised close to $700M in funding in 2018 alone. Conferences around the world are running to packed rooms. Client interest is at an all-time high. There is no doubting the momentum.
We predict that 24 months from now, every major ISV will evolve mature approaches to automation that involve either teaming with RPA/DPA providers or incorporating key features into their platforms. Systems integrators and services vendors will provide a vibrant ecology to supplement product functionality with strong consulting, implementation, and management capabilities. RPA as we know it today will evolve into a broader ecosystem play. The superficial, fragile integration surfaces that exist today will give way to deeper, more robust hooks into enterprise apps.
The Automation We Need Today . . .
Today, RPA is a Band-Aid. Bots cobble together business workflows across a messy patchwork of modern applications and brittle legacy systems.
No doubt, this is what businesses need. Large enterprise IT landscapes have organically evolved over time into a tangled composite of myriad applications. IT staff are often afraid to mess with this tangle because of the simple fact that, despite all the messiness, things work. CFOs love RPA because it allows them to robotize inefficiency and defer capital spending on IT modernization, allowing them to focus on the more pressing transformation projects that they care about.
RPA bots work across a variety of different applications and screens, and they have increasingly started touching enterprise suites such as SAP and Oracle. So it is not surprising that the large ISVs have perked up their ears. Oracle unveiled a hint of a plan toward RPA at OpenWorld earlier this year, but it is SAP that has taken a broader view of bot-based automation across its product suite.
As a first step, it is laudable. The thing is, it isn’t enough.
. . . Is Not The Automation We Deserve (But It’ll Do)
The road to API-driven automation utopia will be arduous and never-ending. Enterprise ecosystems are heterogeneous and fractured, and RPA that largely works inside a walled garden will be of limited value.
RPA works today because it is both pragmatic and opportunistic. Nobody misses the irony of applying cutting-edge modern technology such as computer vision and deep learning to perform tasks as superficial as identifying fields in a legacy screen. However, this kind of irony is precisely what allows AI-enhanced RPA business cases to thrive. Beyond features and functionality, there are broader questions around deployment efficiency, central management processes, and the understanding of how to run a digital workforce. It takes more than bots to make bots work, and RPA product vendors have been heavily investing in capabilities and features that help enterprises manage automation programs effectively at scale.
Now, there is a lot to like about what SAP’s been doing to make its platform open and intelligent (and hopefully more attractive to customers still on the fence about that S/4HANA migration). SAP’s vision for RPA is robust, and it may well be the automation that customers deserve.
However, in the short term, it may not be the automation that customers need.