Every automation project I have been asked to rescue failed in the same place. Somebody bought or built a system for how they believed the work was done. The people doing the work had a different process, usually for good reasons, and carried on doing it their way around the new tool.

Mapping the process first is how you avoid that. It is not a long exercise. For two or three core processes in a business of 20 to 150 people, a week is usually enough.

Talk to the people who do the work

Managers describe the process as it was designed. The people doing it describe the process as it is. You need the second one. Sit with the sales administrator, the warehouse supervisor, the finance clerk. Ask them to walk you through one real example from last week, start to finish, including the workarounds. The workarounds are the map.

Follow one real item end to end

Pick one quote, one order, one client brief, and trace it from the moment it arrived to the moment it was closed. Write down every handover, every system it touched, every time somebody retyped something, and every time it waited. The waits are usually longer than the work.

Mark four things

  • Retyping. Every time information moves from one place to another by hand. Each one is a delay and a source of errors.
  • Bottlenecks. Steps that depend on one person, especially one who travels or is often in meetings.
  • Dark periods. Stretches where nobody can see where the item is or who has it.
  • Decisions. Points where a person makes a judgement. These are the steps you will want to keep human, and design the automation around.

Count before you design

A quote-to-invoice process I mapped recently had seven steps, five friction points and four retypes of the same customer details. Once that is written down, the roadmap almost writes itself: capture the enquiry once, generate everything else from that one record, and keep the sales manager’s approval as the single human decision. The numbers also give you a before-and-after measure, which is how you know the automation worked.

Only then choose the tools

With the map in hand, the tool question becomes much smaller. Often the answer is to connect what you already have. Where something new is needed, you now know exactly what it must do, which makes vendor conversations short.

If you want a second pair of eyes on your map, or would rather someone did the mapping with your team, that is the first stage of how I work.