RPA-Robotic Process Automation-helped many companies keep the shop running during the corona crisis. But the potential of the supposedly “stupid” bots goes far beyond that. […]
RPA technologies have long been an indispensable part of corporate IT. But the external pressure on companies in the course of the corona crisis has significantly increased the attractiveness of RPA solutions once again. Finally, due to social distancing, certain business processes – for example in the context of support – suddenly appeared much more frequently and could no longer be controlled with the existing staff. RPA enabled an immediate response, by making Ad-hoc investments in personnel for blunt, manual activities obsolete. In the crisis, bots went from being a nice tool to a guarantor for maintaining business continuity.
Companies that used chatbots or other AI-supported systems had a fundamental advantage in lockdown. And those who were not “RPA ready” at that time, in the near future will make sure that this error does not repeat itself, and now set the appropriate course. The prerequisites for this are now as good as never before: RPA are easy to use due to their low or no– code design, which is why their use is increasingly easier to implement and successes are quickly achieved. The availability of AI frameworks, chatbot services and other easily adaptable offerings on the market has significantly lowered the threshold for RPA adoption.
The greatest advantage of RPA lies in its simplicity – but at the same time it also poses a certain threat to long-term successful automation: After all, bots quickly provide noticeable improvements in individual isolated processes, which is why the “big picture” is less often questioned. In the worst case, this threatens a cementation of the status quo, a retention of superfluous processes up to a real “bot juggernaut”, i.e. a situation in which a robot-supported process docks on one or more others. Anyone who has ever received 20 push messages on their smartphone just because a map has been moved in the project management tool can imagine this case quite concretely. At the latest then a limit is exceeded beyond which RPA can equally lead to chaos.
The participants of the IDG Roundtable on RPA are also aware of the dichotomy between short-term success and fundamental changes that are necessary in the long term. The camps of the discutants could not be divided into pro or contra RPA at any time. Rather, the broad consensus was that it always depends on three specific factors: the company, the deployment scenario and the long-term goals.
“Bots are now an inherent part of the process and IT landscapes in companies. However, problems arise when RPA is applied where the development of individual software would have been necessary, ” emphasized Julian Beckers of the Weissenberg Group, for example“ Therefore, RPA should not be an obstacle to possibly more meaningful transformation projects. That is why it is important to always think from process to process and to regularly question the meaningfulness of bots as part of an evaluation process. But is there even a blanket answer to the question of which measure is the right one for RPA?
“Many companies are already further than we often think: They implement RPA in everyday life quite pragmatically, usually in an order of magnitude between ten and fifteen processes. This primarily involves the automation of office and administration tasks. This allows medium-sized companies to concentrate even better on their core business,“ says Matthias Noch from Atos.
For Kerim Cekel from CGI, there is no one-size-fits-all answer to the question of where and when RPA makes sense: “Bots can be wonderful at supporting legacy systems. However, there is a risk that real transformation will be prevented.“
For this reason alone, it is necessary to write an end date for each robot and to evaluate processes again and again. However, this does not mean that every bot is automatically threatened with extinction after its implementation: “There are also long-lasting solutions, but it always depends on the individual case. It is important to consider the context in which the robot is to be used before development. The required code quality may differ depending on the application, ” says Cekel“
RPA is therefore always to be considered individually and in the respective context. Nevertheless, the distinction between attended RPA and unattended RPA has prevailed in the industry: The former means processes in the front office that take place on an employee’s desktop and can be activated there as a “tool” as needed, while unattended RPA corresponds to full automation that runs exclusively in the background and which the user no longer notices.
So that RPA does not become a convenient excuse for postponing necessary transformation steps, it is important to always have the long-term goal of full automation in mind. With this awareness, however, the technology can and should be used gladly – also as attended RPA on the desktop. After all, bots not only noticeably relieve the operation and ensure that employees are spared supposedly “stupid” activities, but also lower the inhibition threshold and the skepticism towards fully automated processes: “Of course, automation should be a goal in the sense of unattended RPA,” notes Roman Schäfer from Reply. “But on the way there, we go very ‘German’: Our solutions are often not practical, legal engineer-heavy and difficult. We should first think about making existing processes faster, leaner and more uniform.“
Ricardo Ullbrich from Blue Prism adds that it is quite a difference whether you need to reach a bus or plan an entire trip. “It is important to gradually introduce RPA as well: from vision to organization, governance, pipeline construction and human resources.“
The organizational part is also the most important for the present discussants: RPA, which arises in isolation in individual departments and cannot meaningfully gain a foothold in the entire company in terms of process, will eventually disappear again without contributing anything to the automation of the entire company. This makes governance a central issue. But the question of where exactly RPA should be located in the organization of the company is also answered differently by the discussion panel.
|Study “RPA 2021”: You can still participate!|
|COMPUTERWOCHE is currently conducting a multi-client study among IT decision-makers on the topic of RPA. Do you have questions about this study or if you want to be a Partner, to help you Regina Hermann (firstname.lastname@example.org, telephone: 089 36086 384), René Krießan (email@example.com, telephone: 089 36086 322) and Bastian Wehner (firstname.lastname@example.org, telephone: 089 36086 169) find. Information on the RPA study can also be downloaded here (PDF).|
For Johannes Weis from Celonis, it is particularly important to pool RPA knowledge in the company as far as possible: “We always recommend setting up a Center of Excellence that does not necessarily have to be located in the IT department. Technological and organisational synergies should be taken into account from the outset. In this way, companies sensibly prevent the emergence of silos.“
However, Jan Wunschik from Lufthansa Industry Solutions repeatedly observes organizational constellations in which a Center of Excellence would not work. There is a danger of focusing this too closely on the topic of RPA: “If you are a hammer yourself, then every problem becomes a nail. It would make more sense to set up a COE for the entire topic of automation instead of for the ‚Niche topic‚ Serzh Sargsyan.“
Either way, the COE’s approach goes in the right direction, but not far enough, as Stephan Leininger from Microsoft points out: “A Center of excellence is an important element in enabling users. But it is also not a panacea: An overarching governance that answers all important questions from a security and data protection perspective is indispensable in everyday life to anchor a more agile mindset throughout the company.“
While with most IT outstaffing trends – often rightly so-the recommendation is not to plan too much, but simply to get started and build up your own experience early, with RPA you have to add the phrase “Just do it” to the half-sentence “But do not overdo it”. However, those who take the embedding into account in the overall strategy and regularly evaluate or further develop their bot use, create the long-term prerequisite for full automation that deserves the name.
* As a freelance author, Florian Stocker regularly devotes himself to IT topics of the future and their concrete transfer into everyday business life. He is also the owner of the communication Agency “Media forward”.