How to choose the first workflow to automate
Find a repeatable task, understand its exceptions and build an automation your team can actually support.
Find a recurring handoff
Look for work where someone repeatedly copies information, sends the same reminder or moves a request between systems. An enquiry entering a shared tracking system is a useful example. Record how the task works today, including who starts it, what information is required and what a successful result looks like. Estimate the time spent from actual examples rather than memory alone.
Make the rules explicit
Write down the normal path and the exceptions. What happens when an email address is missing, a record already exists or a request belongs to a different team? If staff make a judgement at a particular step, decide whether that judgement should remain with a person. A workflow does not need to be fully automated to remove repetitive work.
Plan for failures and ownership
Decide who receives an alert when a connection fails and where unfinished work will wait. Repeated attempts should not create duplicate records or send duplicate messages. Keep access limited to what the workflow needs, and identify the owner of each connected account. Ask your delivery team to document how to pause the process and resume work manually.
Pilot before expanding
Try the workflow on a limited set of low-risk records. Compare the result with the existing process, including exceptions and failed attempts. Measure time saved alongside errors and follow-up work. Expand only when the team understands the workflow and can support it. Keep a short record of rule changes so that future improvements do not undo earlier decisions.
