“Robotic process automation requires more intelligence and pragmatism!”

Faster, more cost-effective processes, satisfied employees and customers are what companies can expect from automation with RPA. The reality is usually different, however, as the solutions lack crucial AI components. What are the alternatives for companies that have already implemented RPA and those that are about to do so?

The appeal of robotic process automation (RPA) seems unstoppable – but not for long, if Forrester Research is to be believed. According to the analysts, investments in RPA will increase by a whopping 40% by 2024 but will decline from then on. Why might that be? Has everything that can be automated up to that point been automated? This is unlikely when you look at the processes used by many companies: in IT, production and logistics, RPA is already being used to automate many processes, but the potential is far from exhausted. In HR departments, sales, and management, on the other hand, RPA scarcely been used at all so far. Automation is therefore unlikely to have reached the end of the line in the next three years. However, some companies are likely to experience disillusionment after the initial RPA euphoria. They had hoped for faster processes, better service, and product quality as well as reduced costs. However, all this happened to a much lesser extent than expected, as demonstrated, for example, by the study conducted by IDG and Computerwoche. The not entirely pleasant realization is that RPA solutions can automate processes – but nothing more. So, what now? The good news is that there are interesting alternatives for both newcomers and companies that already use RPA.

Expand RPA intelligently

Let’s assume that RPA has already been rolled out, often in the form of bots that perform several high-level manual tasks. The catch here is that existing processes are mapped 1:1 as they were previously. Transfer data from system A to system B, create a support ticket for every customer chat request: RPA can do this faster than humans, around the clock and with consistent quality. However, this does not improve the process itself. What happens with more complex tasks, for example in customer service? Providing all the customer’s data is no problem, but RPA will not take the appropriate next steps in the event of a complaint. This requires artificial intelligence, for instance in the form of machine learning algorithms that suggest a suitable customer discount in the event of a complaint or, ideally, predict the possibility of the complaint arising even before it is made and automatically initiate the corresponding sales actions.

However, companies that want to complement existing RPA solutions with AI should keep the entire company in mind rather than dwell on individual use cases. Where can other processes be automated and supported by AI? How do they interact, and which solutions can be used across the board to create synergies in implementation and enable cross-process use of all data? An end-to-end strategy and company-wide roadmap can answer these questions and implement automation projects in a sustainable and future-proof manner.

A pragmatic approach to process automation

Companies that are just starting to tackle process automation and want to avoid the RPA trap described above are well advised to go with no-code solutions that include both RPA and AI. Such solutions are offered by Microsoft 365, for example, which most companies already use. Since existing “on-board” tools are used for automation, the new automated processes fit easily into existing applications and workflows, making it easy to get started with process automation. At the same time, employees are already familiar with the applications and have fewer reservations than with completely new solutions. You can even “click” workflows together yourself or use ready-made templates without first asking the IT department for help. Integrated AI components provide support and proactively suggest the next steps. The range of processes that can be automated then spans from simple tasks such as notifications about newly filed documents or processing steps in operations, e.g., in Microsoft Dynamics, to complex multi-stage approval and publication processes, e.g., for customer inquiries or in quality management. The basic idea is that processes are not simply automated, but that employees are supported in the best possible way by smart assistants that perform tasks, even “think for themselves” and thereby free up space for more value-adding activities.


Five theses on Machine Learning Operations (MLOps)

Machine Learning Operations (MLOps) enables companies to efficiently integrate artificial intelligence into every product and service. Just as DevOps has increased the productivity of software […]


From an idea to an automated business process

When it comes to process automation, companies should always ask themselves how they can not only automate their processes, but also streamline them, set them up in the best possible way according to their business logic, and at the same time map them in such a way that they provide employees with ideal support in their day-to-day work. Robotic process automation alone is usually not sufficient for these requirements. However, when combined with artificial intelligence, the result is a powerful solution – whether layered on top of existing RPA tools or embedded, as with Microsoft 365.

However, many companies are only just starting to build up internal RPA & AI expertise. A few strategic questions should provide some initial guidance at this point:

1. Analysis of the status quo: which processes are running well, which are not and why?
2. Target image: what level of automation should be achieved by when, and what specifically should be improved as a result?
3. Roadmap: which specific use cases should be implemented and which of them bring the most added value in the shortest time?
4. Concept: which technical solutions are best suited for implementing the target image and how do they fit into the existing IT landscape?
5. Change & adoption: how can employees be motivated to automate repetitive tasks, and how can they be involved in driving the project forward?
6. Pilot & roll-out: in which area and process should the automation project start? How can lessons be learned and used during the roll-out?
7. Measurement & optimization: at what point can automation be considered successful, where is there a need for optimization and how is this dealt with?

These are just some of the questions that help companies successfully implement automation projects. If you would like to discuss strategies, concepts, and solutions for process automation, please contact Dr. Jens Neuser  via our contact form.