Advantages of Task Mining over Process Mining in process mapping

Generali Česká pojišťovna is a leader in the Czech insurance market. Our priority is to be a lifelong partner and provide excellent service to our clients and partners. To maintain the goodwill of our more than 4 million customers and to be their most reliable partner in the future, we are constantly working to improve our services. This is not possible without continuous data-driven management, very detailed knowledge of processes and the use of modern technologies, such as artificial intelligence.

Therefore, at the end of 2019, we started to address how to map and measure processes correctly to obtain even better data for decision-making. Logically, our first steps were heading toward process mining. However, in addition to the classic process mining, we discovered there is another entirely new option, which was defined by Gartner only last year — task mining. If you, like us, are struggling in your company with how to obtain structured data on the natural functioning of your branches, departments and employees, then our experience could be a good inspiration for you.

The main difference between the two approaches is the definition of the process. While process mining maps and collects data across the company to provide insight into end-to-end processes, task mining focuses on how an employee moves through each application while processing requests, allowing you to see how specific positions and teams are performing in actual operations.

During our PoC, we decided to try both approaches and see what value each can add to companies. For process mining, we chose one of the most widely used tools with thousands of users. Task mining, meanwhile, was represented by a technology startup from the Czech Republic, UltimateSuite, which Gartner named a prime example of this new class of software in its first (and so far only) report on task mining. In both cases, we had a partner implementation team at our disposal.

The fact that these are different perspectives was already evident from the initial requirements at the start of the project.

Process mining needs to be connected to core systems and application databases, then data structures and keys need to be defined to connect the data. If you have many systems in your company and employees use most of them to process one client request, this means you need to examine logs from all these systems and link them to each other unambiguously. Unfortunately, you can't do this without a lot of assistance from the IT department for this whole operation, whether it be from architects or direct analysts and programmers of individual applications. We therefore chose a process for PoC purposes that only goes through one system. But the most significant barrier on our side was that most of the process mining tools needed to move this very sensitive data to the vendor's servers for the analysis phase. This fact also complicates the building of knowledge within the company, increases the dependency on the vendor and delays the time when you can actually start examining the results of your measurements.

Task mining, on the other hand, pulls the information needed for process mapping not from the logs of the core systems but directly from the operating system of a particular machine (Citrix desktop, etc.). In addition, the measurement probe records how an employee traverses each application as it processes requests, so you have near-real-time data from real traffic. It's not entirely jobless, however. There is still the need to define the start and end of the process, which can be challenging given the granularity of the data. The significant advantage, however, is that you only need two servers.

Process Mining vs Task Mining

The outputs are also very different and show that while both categories pull data from internal systems to improve processes, they are very different in scope. For example, task mining provides a very detailed view - the great granularity of data down to the level of the specific employee, the particular task, and screens/buttons in the internal system. But that is not where it ends. It will also give companies a scope of all their teams but focuses primarily on the employee, their work and the employee experience. Process mining, on the other hand, maps the organization. It offers a more systemic, robust view, and the scope of the implementation matches this.

The differences between these solutions, according to our implementation experience, can be broken down this way:

Based on the simplicity of implementation, availability of measured data detail, the flexibility of the solution, and versatility of use, we found task mining to be a better option. We chose the software from the Czech company UltimateSuite, which Gartner also recommends. Task mining’s most significant advantage over process mining was the agility it provides, as UltimateSuite does not need to be integrated into the architecture in any way and can be switched on/off whenever you choose.

The entire measurement project was up and running in a matter of days. During the first week, we were getting exciting insights from the measurements, showing us, for example, the differences in the way employees perform the same tasks or seeing in the data where they spend the most time in the process and which part makes the most sense to focus on.

Task mining offered the management of Generali Česká pojišt'ovna new insights into the processes of handling client requests. The measurement was conducted across teams, collecting all requests and bringing in a large volume of data. The conclusions from the measurement were much more conclusive and robust than with other methods.

That's why we decided to deploy task mining for real at Generali Česká pojišt'ovna and map all the agendas in the back office (communication center, claims, insurance administration). It gave the company management a concrete output with recommendations on what to focus on and improve further. This includes recommended practices in retention calls, answering client emails, handling company chat, indexing documents, etc.

Thanks to the level of detail that the new data provides, it is easy for team leaders and managers to identify best practices and bottlenecks and potential causes of inefficiency. It is a beneficial complement to the other statistics from internal systems that we use daily to manage our teams. What used to be detective work is now almost automated insights about team performance. Managers and coaches have input on improving the processes for which they are responsible.They also now have information on how to best help their colleagues — organizing training, adapting methodological materials, or improving the knowledge base, for example. At the same time, the aim is to encourage colleagues to share know-how. We also want to use the results from the measurements as input to define a new frontend in the communication center and disposal.

Another area where it is possible to use the collected data is enterprise architecture. IT can use the output from UltimateSuite as a basis for examining the application portfolio. They can find out which specific applications are being used by which teams and how often, including an overview of the usability of particular screens or buttons. This information is valuable if you are thinking about redesigning an existing system or building a new system. For example, we found through task mining that the email team operators were using their local templates, instead of the templates in the internal system to reply to the customers. It helped us in prioritizing and then implementing a more suitable solution.

At Generali Česká pojišťovna, we want to provide excellent service to our clients and partners, which means constantly innovating. As you can probably easily conclude from my comparison, task mining makes a lot of sense for us. While traditional process mining gives value in more extensive infrastructure mappings(for example, for RPA), UltimateSuite has offered us an agile tool to quickly gain insight into how processes work in a natural environment and how specific employees work to perform them. We see the advantage of working specifically with UltimateSuite mainly in the flexibility of developing the entire solution and the service that even this smaller company can provide.