How to Choose the First Workflow to Automate

Begin with the work people experience repeatedly. Ask where requests wait, where details are re-entered, where follow-ups depend on memory and where clients experience avoidable uncertainty. Frequency matters because a small inconvenience repeated every day may cost more than a dramatic problem that happens twice a year.

Avoid choosing a workflow merely because it will produce an impressive demonstration. The first project should improve an outcome the business values, such as response time, completeness, visibility or consistency. A clear operational benefit makes adoption easier because the team understands why the change exists.

• Prioritise frequent and observable friction
• Tie the workflow to a business outcome
• Avoid novelty as the main selection criterion

A simple scoring exercise can bring discipline to the decision. Rate each candidate workflow by frequency, time consumed, error exposure, customer impact and reporting value. Then consider complexity, exception volume, data sensitivity and the cost of failure. A high-impact workflow with uncontrolled risk is not automatically the best starting point.

The strongest first candidates are usually repetitive enough to define but not so rigid that every unusual case breaks the process. They have a routine path, a limited number of known exceptions and a responsible person who can explain how decisions are made. This creates space to learn while protecting important business controls.

• Compare opportunities using shared criteria
• Include both value and operational risk
• Prefer workflows with known exceptions

A workflow needs a clear starting event and a clear completion point. Automate new enquiry capture to sales assignment, for example, rather than attempting to automate the entire client relationship in one project. Boundaries keep discovery focused and make it possible to test whether the workflow performs as intended.

Ownership is equally important. The process owner should understand the current work, have authority to clarify decisions and remain accountable after launch. Automation does not remove process ownership. It makes good ownership more visible because rules, exceptions and responsibilities must be stated rather than assumed.

• Define the trigger and completion event
• Keep the first scope coherent and contained
• Name an accountable process owner

A first workflow should have a baseline and a small set of meaningful measures. Record the current turnaround time, correction rate, number of follow-ups or percentage of requests with complete information. The baseline does not need to be perfect. It needs to be consistent enough to show whether the change helped.

Also define what acceptable failure looks like. Decide how the team will see an exception, who will intervene and how work can continue if an automated step is unavailable. This is part of success, not a separate technical concern. A dependable workflow is one the business can understand and control when conditions change.

• Capture a practical before-state
• Select measures linked to the intended outcome
• Plan exception handling and continuity