Task mining: lightning-fast set-up and use. What’s the secret?
We still don’t find it so easy to find information to help you set up and run a successful task mining project within your organization. So in this article we aim to clarify the benefits of a task mining platform in order to allay any misunderstandings about how the underlying technology works, how quickly and easily it can be set up, and also provide insight into its full capabilities.
The experience of Generali Insurance is pretty typical, and will certainly resonate with many other companies that have been considering task mining to support operational excellence initiatives.
Generali is a leading insurance company with 4 million customers and several thousand employees. A few years ago they recognized the need to protect and increase their customer base through a better end-user experience. This required streamlining business processes and improving overall operations. They quickly reached the conclusion that existing legacy tools and methodologies could not provide the insights required to accurately map the current state of their operations, and so they looked into process mining and task mining.
After running a simultaneous PoC with the two technologies it became apparent that process mining had certain shortcomings. Firstly, because it is, by definition, an intrusive technology which requires integration with all core systems and databases from which it captures log files, and therefore needs heavy involvement from IT to get set up, which inevitably slows things down. Secondly, although a process mining solution can provide a comprehensive overview of the organization’s processes, it does not identify the use of productivity tools involved in the thousands of workarounds that evolve over time (let’s not pretend to be surprised, we all know it happens). However, the most worrying shortcoming is that most process mining solutions are obliged to process highly sensitive data off-site. That’s a big red flag for nearly everyone.
Generali realized that task mining did not necessarily have these shortcomings because it is normally deployed on-premise, which brings a host of advantages with it: it required no integration with core systems (and so no painfully slow IT intervention), so implementation was more or less instantaneous; the software automatically captured all steps taken, including both core systems and office productivity applications, making it highly granular, and consequently provided detailed insights into the efficiency of specific tasks within processes. As this was raw data processed on-site, these insights were available within minutes, rather than days or weeks. And finally with regards to data sensibility, two things: first, they noted that the activation of task mining can be controlled by the individual employee performing the task, thus preserving their intimacy; and second and most important, this sensitive data never leaves the four walls of the organization, mitigating any data security risks.
Ultimately, task mining provided the company management with a concrete output including recommendations on what to focus on and improve further.
What used to be frustrating detective work is now almost fully automated. They also realized further benefits of task mining, one being that their IT department could use the captured output to examine their application portfolio in fine detail and make informed decisions about application retirement. More savings were found in this area.
In simple terms a task mining platform has the following elements:
An in-house server for processing captured data set-up with the assistance from an implementation team. The process is fast, usually being completed within hours, not days or weeks.
A software agent is deployed on all user workstations from which you wish to capture data. An implementation pack is supplied. It is recommended to deploy the agents to all user desktops throughout an organization. This is an automated process, enabling multiple projects to be set-up and run with minimum lead times across various departments. Users can switch the agent on and off at will.
Access to a BI platform for viewing the processed data.
