Task mining to generate enterprise value

The convergence of process and task mining within a wider process intelligence market

From being a purely academic discipline back in 2011, process mining has evolved significantly over the last decade and has become a catalyst for digital transformation and enhanced process efficiency within many corporations. The increased investments in process mining and the associated development of a related multi-billion software market are also evident in the inaugural publication of the Magic Quadrant^TM for Process Mining Tools by Gartner in March 2023. Various acquisitions and partnerships as well as the latest product portfolio updates herald even more disruption on the horizon with the imminent convergence of process mining and task mining.

So far, our blog series has focused purely on process mining and outlined its value as well as describing our proven deployment approach, following the Focus, Act and Scale phases. This article turns to the concept of task mining and presents its main areas of application and use cases.

What is task mining and how does it differ from process mining?

Task mining relies on recorded desktop activities (i.e., front office systems such as Excel, Outlook, etc.) complemented by machine learning algorithms to identify sequences of activities completed by single users or teams. Activities being collected within task mining may range from simple keystrokes and mouse clicks to more complex user interaction patterns involving inputs across various front office applications.

As already pointed out in our previous blogs, process mining relies on event log data from back-end systems (i.e., ERP systems, CRM or PLM systems, etc.) to model and construct end-to-end processes by connecting case IDs with specific time stamps.

In a nutshell, both methods zoom to different granularity levels and provide different types of insights. Task mining provides a picture of the behaviour at the individual or team level, whereas process mining allows for insights across overarching, enterprise-wide business processes. Additional differences and similarities between task and process mining are illustrated in Figure 1.

What are the main use cases of task mining?

Like the main stages of end-to-end process transformation, the major use cases of task mining relate to discovering, optimising, and automating tasks. Within task discovery the aim is to assess how certain tasks are being performed by individuals or teams and to benchmark against the expected outcome. When identifying any deviations or means of making improvements, task optimisation can be activated and embedded within an overall continuous improvement roadmap. We increasingly see organisations not only focusing on the identification of peak workloads and the associated staffing optimisation but also on determining if appropriate times are spent within certain applications. If the analysis outcome points to automation potential, task automation can be included into the overall automation agenda to generate value.

Given current macroeconomic conditions, we see a significant uplift in demand for task mining within overall productivity and/or cost transformation programmes. Most organisations also leverage the use of task mining to detect evidence-based signals for the continuous improvement of overall team engagement, collaboration, integration, and wider employee well-being. To make these enterprise-wide efforts a success, some key factors need to be addressed, with a particular focus on data privacy, which will be outlined more thoroughly in the next section.

What are the key strategic considerations associated with task mining?

In line with the presentation in Figure 1, key strategic considerations for employing task mining encompass data privacy topics, how to drive change management and adoption across all employees as well as how to effectively use and scale task mining given its enhanced time-to-value when compared to process mining.

As task mining relies on recording activities performed by individual users or teams, data privacy is crucial. Anonymising or pseudonymising critical data represents a valuable starting point along with restricting data via data pool permissions and performing analyses for employee groups to preserve confidentiality. While most of these measures are already out-of-the-box functionalities in leading task mining tools, they might not be sufficient in countries with strict data privacy regimes, where alignment with workers councils becomes a critical factor. As a result, early engagement and involvement with data privacy experts and workers councils will be essential to the success of the task mining programme, along with proper internal communication and change management.

While supporting various task mining initiatives we have seen all major use cases of task discovery, task optimisation and task automation being deployed effectively once they have been aligned with the overall strategic agenda and have acquired proper executive buy-in. The scope of task mining projects might vary from initial proof-of-concepts and individual proof-of-value initiatives to wider incorporation into the overall automation and digitalisation agenda. In addition, we observe that leading organisations already making use of process mining at scale (characterised by a dedicated Centre of Excellence and wide-scale adoption of process mining across the company) converge into task mining to complement the end-to-end process insights with more granular information at the individual employee level.

How task mining can complement process mining: an Order-to-Cash example

The combination of process and task mining provides unique insights from manual activities that occur outside key systems, providing a 360°-degree view across both back end and front-end systems. To prove the value of combined process and task mining, we supported a leading consumer goods company as they further refined the initial analysis of their Order-to-Cash process (conducted via process mining) and identified and implemented further improvement opportunities.

As shown in Figure 2, the assessment revealed that most of the time (analysed on Level 1 Order-to-Cash process steps) is spent outside the SAP system, which is the client’s main ERP system and the major system employed across the Order-to-Cash process. As a result, we deep-dived into selected areas and teams across the overall process, focusing on the transitions between all applications used by employees working across the process. Through the application of task mining, it was therefore possible to refine the overall effort estimates for the baseline FTEs and identify exceptions and the “Happy Path” process steps based on volume (process mining) and time spent (task mining).

The combination of process and task mining enabled the definition of more detailed continuous improvement initiatives, pointing to a 40% reduction of the total effort by eliminating exceptions and automating the “Happy Path”. This was accomplished as task mining brought enhanced visibility to the remaining parts of the Order-to-Cash process. Only 21% of the time was indeed spent in SAP vs. 54% spent in the Microsoft technology stack. Overall, the combined approach involving task and process mining showed promising potential to be scaled across other process areas within the client’s organisation, providing more granular insights and enhanced improvement potential.