.jpg)
Process mining programmes often begin with technology: which platform to use, which systems to connect, and how quickly the first process model can be produced. Yet one of the more consequential decisions comes earlier. Organisations must decide which process to analyse first.
This is not always obvious. A large enterprise may have hundreds of processes spanning finance, procurement, supply chain, customer service and operations. Many will produce the event data required for process mining, but technical feasibility alone does not make them equally valuable candidates.
Research presented at the European Conference on Information Systems in 2022 identified 20 criteria for assessing process mining use cases, covering factors such as business importance, process weaknesses, data quality, organisational support and optimisation potential.
For enterprises, the practical implication is straightforward: process mining should begin where there is a credible combination of business value, usable data and an ability to act on the findings.
A useful assessment begins with the outcome the organisation wants to improve. This could include reducing cycle times or rework, improving working capital, strengthening compliance, increasing throughput or addressing a persistent service problem.
The use case should be specific enough to connect process behaviour to that outcome. “Analyse accounts payable”, for example, is too broad to provide much direction. A better objective would be to identify which exceptions contribute to invoice-processing delays and where avoidable rework occurs.
This distinction matters because process mining is most useful when it is applied to a business problem rather than treated as an exercise in process visualisation.
Business importance needs to be considered alongside data readiness. Process mining reconstructs process execution from event data recorded in enterprise systems, which means the quality and accessibility of that data can determine whether an otherwise promising use case is practical.
Organisations should establish whether process instances can be identified consistently, whether activities and timestamps are reliable, which systems contain the relevant events and whether those systems can be accessed.
The ECIS research on process mining use-case selection places particular emphasis on data availability and quality. Its case analysis also illustrates the trade-off: processes with high business importance may still be difficult starting points when their data is incomplete or difficult to interpret.
For an initial project, a slightly less ambitious process with dependable data may produce a stronger foundation than a strategically important process whose data requires extensive remediation.
High-volume processes are often attractive because repeated execution provides enough activity to identify recurring variants, delays and exceptions. Order-to-cash, procure-to-pay and accounts payable are common examples.
Volume, however, should not be considered in isolation. A lower-volume process may deserve attention if it handles high-value transactions, carries substantial compliance exposure or supports an important strategic initiative.
The same applies to enterprise transformation. An organisation preparing for an ERP migration may prioritise a process because it needs evidence of how work is currently performed before deciding what should be standardised or redesigned.
The objective is therefore to understand the significance of the process within the wider operating model, rather than simply choosing whichever process produces the most transactions.
Actionability is another important consideration. Process mining may expose unnecessary hand-offs, recurring exceptions, excessive approvals or process variants that take considerably longer than others. Those findings have limited value if the organisation lacks the authority, resources or appetite to address them.
Automation potential should be assessed in the same way. A process containing repetitive manual work may be a strong candidate for automation, but analysing the process first can reveal whether that work should be automated at all. In some cases, redesigning or removing an unnecessary step will make more sense than automating it.
This becomes increasingly relevant as organisations explore intelligent automation and AI agents. Better automation decisions depend on a reliable understanding of the process into which the technology will be introduced.
Verdant Data recommends assessing potential use cases across six dimensions: business value, data readiness, process volume, strategic importance, risk and automation potential.
Each candidate can be scored from one to five across these dimensions to create a consistent basis for comparison. The score should inform management discussion rather than dictate the final decision. Weak data readiness, for instance, may make a high-scoring process unsuitable for an initial deployment, while significant regulatory exposure may justify prioritising a lower-volume process.
This approach also helps bring the relevant stakeholders into the same conversation. Technology teams understand what data can be accessed, business teams understand which outcomes matter, and process owners understand which changes are realistic.
The strongest process mining use cases tend to combine three characteristics: there is a material business problem, sufficient evidence exists to investigate it, and the organisation can respond to what the analysis reveals.
That provides a more useful starting point than asking which process can be connected to a process mining platform most quickly.
Process mining creates greater value when organisations begin with a business question and use process data to investigate it. Once that foundation is established, the same capability can support broader process intelligence, continuous monitoring and more informed decisions about transformation and automation.
─────────────────────────────────────────────────
There is no universally best process. A strong starting point combines measurable business value, usable event data, sufficient process activity and a realistic opportunity to act on the findings. Research on process mining adoption similarly suggests that multiple organisational, process and technical criteria should inform use-case selection.
Traditional process mining generally requires event data that can identify a process instance, the activity performed, and when it occurred, commonly represented by a case ID, activity, and timestamp. Additional attributes can provide richer analytical context.
Processes supported by digital systems and recorded event data are potential candidates. Enterprise applications span finance, procurement, customer service, supply chain, manufacturing, and other operational domains. Research also shows process mining supporting objectives including transparency, efficiency, quality, compliance and agility.
Process mining can be particularly useful before major automation decisions because it provides evidence about actual process execution, variants and exceptions. Research into the intersection of process mining and RPA has specifically examined process mining as a way to support automation discovery and implementation.
A structured scorecard can compare candidates across business value, data readiness, volume, strategic relevance, risk, and automation potential. The weighting should reflect the organisation's objectives rather than treating the score as a universal formula.
Get Updates and announcements from the Verdant Data Team
You can unsubscribe any time. Learn more about our Privacy Policy
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
%20(1).jpg)
.jpg)
.jpg)
.png)
