This article was originally written for RPA Journal: https://rpa-journal.org/fragmented-automation-implementations/
RPA has achieved widespread adoption to a large extent owing to its promise of quick and decentralized process automation. However, high velocity of implementation and proximity to process owners has its cost in inevitable proliferation of RPA platforms, systems and approaches.
As a result, large enterprises end up with disparate “islands” of automation, based on different technologies, guided by different leaders and supported by different partners. Up to a certain threshold it does not constitute a challenge. However, as the enterprise continues to scale its automation initiative, the costs start to add up.
After starting its automation initiative in 2015, Deutsche Telekom quickly scaled to over 450 processes and 3,000 bots in 2019, becoming one of the largest users of RPA in Europe. At this point in time, the company operated seven different RPA platforms and was struggling with the amount of change requests resulting from any adjustment of system landscape or UI, as well as the workload associated with maintaining bots. The decision was made to reduce the number of RPA platforms down to three and to optimize their code within the scope of a broader business process management initiative (whole story can be found here).
Deutsche Telekom is not alone in its experience. In a 2020 paper, Gartner explained how many of its clients report eight or more automation tools that are being deployed and utilized in silos.
Besides accumulating technology debt, arguments for consolidating RPA landscapes include the benefits of unified process orchestration, simplified application landscape, and achieving economies of scale.
Unified process orchestration: having a single command center over organization’s processes helps reduce complexity, increase resiliency, improve time to service ticket resolution, and simplify audit. Process automation efforts have also revealed the potential for standards-based process optimization and unification.
Reducing the number of applications: while organizations are starting to rationalize their IT landscapes, they are still managing hundreds of applications. According to a recent MuleSoft survey, the average organization runs 843 individual applications (down from 1,020 in 2018), contributing to 68% of IT time being spent on maintenance activities. Using disparate RPA technologies entails utilizing their respective automation ecosystems, multiplying the number of applications in use.
Scale effects manifest themselves in several ways: better bot pricing from RPA vendors and implementation partners, higher reusability of standard code snippets saves development time, less time required for CoE developer training on various platforms.
Axel Springer has been using a single RPA vendor. “We are only using one solution (Automation Anywhere). It was a conscious decision to go with one platform; so far it has proven very good for us. If we had to expand it in the future, it would likely be PowerAutomate Desktop, because we already work with PowerAutomate” – comments Benedikt Böhme, Senior Consultant and ex-Head of RPA at Axel Springer.
On the other hand, there are arguments against forceful consolidation of the company’s RPA landscape.
Most importantly, sometimes there are good reasons why business units choose different platforms: functionality fit to the business unit’s needs or availability of better support from a trusted implementation partner. Transition to another platform could mean the use of important functionality or know-how for a local business unit.
Another commonly cited factor is vendor dependence. RPA industry is undergoing dynamic changes, illustrated by the recently announced acquisition of Blue Prism by private equity firm Vista, with the plan to integrate it with TIBCO. Some might be worried about the future direction of one or another vendor and would like to maintain a hedge. However, it is arguable whether using different RPA vendors really reduces the risk of vendor dependency, as long as there is no straightforward way of bot migration between proprietary platforms.
There are approaches to enable companies to use the best of both worlds, such as the Choiceworx dashboard for monitoring and remediating bots from various vendors. Such professional solutions might replace home-grown RPA monitoring dashboards, enabling companies to have greater visibility into their automation landscapes. Still, these solutions might be too narrow to address the dilemma presented in this article.
As companies’ RPA initiatives mature and they consider the benefits of reducing the number of vendors, competition will intensify to win over enterprise clients, contributing to further consolidation in the RPA market over the coming decade.