Not every cumbersome workflow needs to be digitized immediately. A strong first process combines visible friction with a clear outcome and a scope the team can actually manage. This guide explains how to recognize improvement potential in daily work, assess consequences rather than symptoms and choose a starting point that produces reliable learning quickly.
Observe friction where the work really happens
Look for recurring situations in which people copy information, ask for the current status or maintain private side lists. Long approvals, ambiguous handoffs and customers who have to follow up repeatedly are also strong signals. Collect concrete examples over several days or weeks. A single complaint may be incidental; a repeated workaround shows that the workflow itself needs better support.
- Record duplicate entry and parallel lists
- Observe waiting time and repeated follow-up questions
- Mark unclear ownership at handoffs
Assess the effect on teams, customers and decisions
Consider more than how irritating a task feels. Does missing context lead to a poor decision? Does an approval delay delivery? Does a customer lose confidence because a commitment cannot be found? Time matters, but so do quality, risk, customer experience and dependence on individual knowledge. A low-volume process can be more important than a frequent one when the consequence of an error is particularly serious.
- Review time, quality and error consequences separately
- Include effects on customers and other teams
- Make dependence on personal knowledge visible
Prioritize value, urgency and feasibility together
Compare candidate workflows using a small number of clear criteria: how much friction exists, how valuable a better process would be and how clearly roles and outcomes can be bounded. A good first candidate has recognizable value, available process owners and manageable dependencies. Do not automatically choose the largest or most visible workflow. A smaller process can create a stronger foundation when the team can learn faster and establish the change safely.
- Assess value and urgency transparently
- Consider whether process owners are available
- Evaluate dependencies and data realistically
Turn the problem into a clear first use case
Describe the starting point as one connected workflow with a trigger, participants and desired end. “Track an enquiry from arrival to a confirmed response” can be tested; “improve customer service” cannot. Decide which information is captured centrally, who takes the next action and which statuses genuinely support a decision. Automation comes afterwards: it should accelerate a workflow the team understands, not move ambiguity around more quickly.
- Name the trigger and successful end clearly
- Define participants and next ownership
- Plan automation only after the workflow is clear
Questions about this topic
Questions about this topic
Do we already need precise metrics?+
No. Concrete examples, observed waiting time and recurring questions are enough for an initial assessment. Metrics can then help track the starting point and later change more systematically.
Should we begin with the largest process?+
Not automatically. The strongest starting point combines relevant value with a clear scope, available ownership and manageable dependencies.
Does every step need to be automated?+
No. Clear ownership, a shared status and dependable data often improve a workflow substantially. Automation is most useful where a repeatable rule is already understood.