Guide · Automation
Process mapping before automation
Automation makes a process faster — including its flaws. A process with unnecessary steps, unclear responsibilities and hidden exceptions becomes a fast, confusing process once automated. Mapping the process first — as it really works, not as the instructions say — often reveals that some steps can simply be removed, which makes automation cheaper and more reliable.
[ 01 ]
In short
In short
- 01
Map the process as it is today, with the people who do the work.
- 02
Capture steps, handoffs, systems, data, waiting times and exceptions.
- 03
Remove or simplify before automating: duplicate checks, habit steps, unnecessary approvals.
- 04
Define rules explicitly, including what happens in exceptions.
- 05
Automate the stable, repetitive parts first; keep human judgement where it adds value.
[ 02 ]
Guide
Step 1: choose the process and its boundaries
Pick one process with a clear start and end: from enquiry to proposal, from order to delivery, from purchase request to payment. Define what triggers it and what counts as done.
Step 2: talk to the people who do it
Interview each role involved. Ask them to walk through recent real cases, including difficult ones. Formal procedures often differ from practice; the differences are where automation fails if missed.
Step 3: map the steps
For each step, record:
- who does it;
- what triggers it;
- what information is needed and where it comes from;
- which system or tool is used;
- what is produced and where it goes;
- how long it takes, including waiting;
- what can go wrong.
A simple swimlane diagram — one lane per role — makes handoffs visible.
Step 4: find the waste
Look for:
- duplicate data entry: the same information typed into several systems;
- habit steps: checks or reports nobody uses;
- waiting: approvals that sit for days;
- unclear ownership: steps where nobody is responsible;
- workarounds: spreadsheets and chats compensating for missing functions;
- rework: errors caught late and corrected.
Step 5: simplify
Before automating, remove unnecessary steps, merge duplicates, clarify responsibilities and reduce approvals to those that matter. This often delivers a large part of the improvement with no technology at all.
Step 6: define rules
Automation needs explicit rules: "Orders over a certain amount require manager approval", "Enquiries in Russian go to team B", "If payment is not received within a set time, cancel the reservation". Write down every rule, including exceptions and what happens when data is missing.
Step 7: choose what to automate
Good candidates are frequent, repetitive, rule-based steps with structured data: data transfer between systems, notifications, document generation, status updates, reminders. Keep people for judgement, negotiation, exceptions and customer relationships. See workflow automation.
Step 8: measure before and after
Record baseline metrics — time per case, errors, waiting time, cost — so the effect of changes can be measured.
[ 03 ]
FAQ
Frequently asked questions
How detailed should the map be?
Detailed enough to show handoffs, data and exceptions. Very detailed maps of rarely performed steps are not worth the effort.
Who should own the process after automation?
A named person responsible for the process outcome, who approves changes to rules and receives alerts when automation fails.
Can we automate without mapping?
For a very small, obvious task, yes. For processes involving several people or systems, skipping mapping usually leads to rework.