Skip to content
Chroma Div

What to automate first in your supply chain: a 5-question test

By Sohaib Latif, Lead Supply Chain Technologist. . 5 min read.

The short answer: Automate work that is frequent, rule-based, already has data, is costly when it goes wrong and has a clear owner.

The five-question test

Most automation projects that disappoint were pointed at the wrong task. Before choosing software, score each candidate task against five questions.

  • Does it happen at least weekly? Rare tasks do not repay the build.
  • Could you write the rules on one page? If two experienced people would decide differently, fix the process first.
  • Is the data already captured somewhere, even in a spreadsheet?
  • Is a mistake or delay expensive: a stock-out, a late shipment, an expedite fee?
  • Does one person own the outcome and want the change?

A task that gets five yes answers is a good first project. Three or fewer, and it is better to tidy the process before automating it.

Four tasks that usually pay back first

  • Weekly reporting. If someone spends a day joining exports for the Monday meeting, a live dashboard returns that day every week.
  • Purchase order follow-ups. Automatic reminders and a view of late orders replace the phone and email chase.
  • Reorder suggestions. Reorder points recalculated from current demand cut both stock-outs and excess stock.
  • Document entry. Invoices, quotes and drawings read into structured data instead of being retyped.

None of these need a new ERP. They sit on top of the systems and spreadsheets you already run.

When not to automate yet

  • The process changes every month. Automating it will lock in a version nobody keeps.
  • Nobody trusts the master data. Fix item codes, units and supplier records first.
  • The goal is headcount reduction alone. The reliable gains come from fewer errors, faster decisions and time returned to planning.

How to estimate what it’s worth

Multiply the hours spent each week by a loaded hourly cost and by 48 working weeks. Then add the cost of the errors the task causes today: expedited freight, lost orders and excess stock.

If the yearly figure is a multiple of the build cost, start there. If you are unsure of the numbers, a 30-minute audit is usually enough to size the first project.