Thanks to Macros Reply the insurance company Helvetia Austria is successfully using Robotic Process Automation – in short RPA – in order to offer customers significantly reduced processing times. The use of RPA allows processes to be automated efficiently and without extensive interface developments. New, powerful tools from companies such as UiPath make RPA increasingly attractive even for smaller cases. The pressure to digitize is weighing heavily on insurance companies. Customers demand faster services. However, inflexible legacy systems make it difficult for IT departments to react quickly to these requirements. In this environment, Robotic Process Automation (RPA) opens up new opportunities. Whereas in classic automation, cross-system processes require lengthy interface developments, RPA software-bots comprehend what administrative staff do manually in terms of work steps, comparisons, entries and data transfers. This enables companies to efficiently automate complex, cross-system work loads.
"It is a paradox: we and other insurers employ highly qualified specialists for work where often there is nothing to decide. Many simple processes could be perfectly automated. Efficient solutions are needed here. Otherwise it is frustrating from the specialist departments’ point of view: routine work for highly qualified employees, and customers who sometimes wait incomprehensibly long for answers and notifications. Our RPA projects focus on the goal of significantly improving the customer experience," says Gerald Klammer, Head of Insurance Claims at Helvetia Austria. RPA is not new from a technological perspective. Many insurance companies already use this technology. As a result, RPA tools have made enormous progress in recent years. Whereas bots were previously only used in high-volume and relatively simple processes, the use of RPA is now also worthwhile for the automation of processes with medium case numbers and numerous process variations as well as in more complex IT landscapes. This progress brought Helvetia Austria into contact with the software specialist Macros Reply. The consultants have process expertise from the insurance industry which they combine with software from UiPath's leading RPA tools – according to Forrester Research and Gartner. "The idea of RPA with UiPath is to use pilot projects to familiarize customers' specialist departments with the tool and the way bots work, and thus win them over to automation. The creation of bots is modularly structured at UiPath. This means that proven functions can be easily copied into other processes and bots at a later date. This reduces effort and increases quality at the same time," reports Ralf Scheuchl, CEO of Macros Reply in Munich.
Together with Macros Reply, a team around Helvetia’s project manager Stefan Hackl initially started an analysis with the help of Process Mining. Thereby, software-supported processes based on historical data, and their variations were examined. In this way, KPIs for runtimes, idle times, net and gross times as well as the number of variations for various processes can be quickly determined. Results from process mining reflect the reality and are much more informative than the process descriptions available in insurance companies. The results were then discussed in a workshop. "This analysis was an important and very profitable step. I can only recommend it. We were able to slim down and optimize the processes even more before automation. On the other hand, this also allows us to identify processes that are particularly suitable for RPA in terms of case numbers and complexity – and which have a great leverage effect on the customer experience," says Stefan Hackl. The process mining revealed enormous process spreads. For example, in a process of a legal advice case, the runtimes ranged from 22 seconds to almost 7 days. The average was 1 day and 6 hours – for an average damage amount of around 96 euros. Another process identified as RPA-compatible included the purely manual transfer of an external preliminary damage report into the insurance policy. Here, RPA was able to enormously simplify the data transfer between different internal and external systems without having to program interfaces. As a concrete RPA pilot project, the team finally chose the processing of glass damage. The decisive test step for the accounting clerk here is the comparison with the premium payments: Once a policyholder has paid his premiums, payment is made – but in between there are many manual entries and data transfers. According to Process Mining, the average processing time in the specialist department was – with a wide range of fluctuations – 1 day and 9 hours, with a net processing time of only 6 minutes. So, there was a lot of potential for a better customer experience.
In the first step, a process design document was jointly developed, which described the technically desired process variants referred to as the "happy path”. Based on agile methods, Peter Preuss, RPA specialist at Macros Reply, further developed this document during the bot creation process. A first important finding: the modular structure of bots with UiPath proved its worth. Individual work steps could be discussed, programmed, tested as partial functions and then used again and again in the bot. "For me, this modularization of the bot that is possible with UiPath is a central quality feature. This is the only way to make the way the bot works really transparent and the individual functions and tasks can be meaningfully tested and reused in other bots. This high quality then generates acceptance in the specialist departments," says Peter Preuss, giving an insight into his experience with RPA projects.
UiPath offers another option for winning over employees in claims processing for bots. A bot can be operated in two modes – attended and unattended. Unattended, the bot starts all processes automatically and then reports the successful work. In attended mode, on the other hand, a bot is specifically started by the administrator, who observes the individual steps such as calls and data transfers. This mode of operation is particularly useful in the test and introduction phase because the clerks gain confidence in the technology while monitoring the way it works. In addition, sub-steps of processes can also be automated in this way via RPA, which can be started by the administrator as required. Regardless of the mode, if a bot encounters difficulties or irregularities, processing is aborted and escalated to a clerk. Within the Macros Reply solutions, this is done, for example, in a process-secure manner via the inbox system. This creates security and trust. In addition, the escalation allows only in the first step the most frequent or simplest cases to be covered by a bot. The case coverage can then be successively increased, which corresponds to an agile working method that supports the rapid deployment of bots even in processes with many variants. Today, at Helvetia, the processing of glass damage is completely handled by a bot – acceptance, data transfer, inspection, communication and payment instructions. What the clerks are pleased about: They are relieved of simple tasks that are perceived as particularly monotonous. The plus for the claims department: instead of only checking in random samples, Helvetia now checks the authorisation via the premium receipt for each settlement request – with reduced processing costs. "The most important thing, however, is that our customers benefit from significantly reduced processing times", says Stefan Hackl happily.
The Helvetia team around Stefan Hackl evaluates the experiences from the RPA project and the help from Macros Reply very positively. The tools of UiPath have indeed proven to be easy to use. Helvetia Austria has already initiated further RPA projects with Macros Reply. To anchor this organizationally, Stefan Hackl wants to concentrate the bot creation in a special team in IT. "With a well-established team that uses the RPA tools on a permanent basis, the efficiency and quality of the bots can be better guaranteed than with changing project teams. This team will involve the colleagues in the Claims Service in the bot creation process. There will also be an exchange of experience between the individual departments every two weeks to ensure the transfer of know-how and to initiate new projects. During development, the IT and business departments will work closely and continuously on the bot solution in order to quickly resolve problems and errors". The test phase of the glass damage bot shows how quickly errors are corrected. Here, at first, breakdowns appeared, but no deviations from the process or case coverage could be found in the aborted cases. The cause was found elsewhere: the response times (latencies) from the core systems fluctuated greatly. Where the human administrator showed patience and simply waited, the bot showed much more impatience and aborted processing. Once detected, the bot was also taught the necessary patience. Acceptance in the specialist department increased quickly as a result. At Helvetia, the people and departments involved are highly satisfied with RPA. With comparatively little effort and completely without complex interface developments, processes could be automated and throughput times were considerably reduced and stabilised. Customers are thus offered a significantly improved service experience. At the same time, the employees feel relieved in the right places and of the right tasks.