How to identify automation opportunities in your operations.

Most operations teams already know which parts of their day are wasteful. The problem is not identification — it is prioritization. Not every manual step is worth automating, and picking the wrong first target burns budget and goodwill.

Here is the framework we use in discovery sessions. It takes about an hour with a small group of operators and produces a ranked list your leadership can act on.

Question 1 — Where does the same information get retyped?

Ask your team where they copy data from one system into another. A field from an email into a CRM. A number from a PDF into a spreadsheet. A status from a WhatsApp message into an ERP. Retyping is the clearest signal of missing integration or missing extraction — and it is almost always automatable.

Question 2 — Which decisions get delayed waiting for information from another team?

Every operations team has a version of this: 'I can't sign off on X until Y sends me Z.' Sometimes the fix is process; often the fix is a system that surfaces the missing information the moment it exists, instead of when someone remembers to send it.

Track these dependencies for a week. The ones that recur are the ones worth automating first.

Question 3 — What does your best operator do that no one else can?

In every team there is one person who 'just knows' — how to categorize an unusual invoice, which exception to escalate, which supplier to call. That tacit judgment is a bottleneck and a risk. When they leave, the process breaks.

Intelligent systems are good at capturing that judgment as decision logic — not perfectly, but consistently enough that the rest of the team can operate at 80% of the expert level without the expert in the loop.

How to rank the list

For each candidate, score three things: frequency (how often does it happen?), cost per instance (minutes of skilled time), and risk if wrong (small annoyance or material problem?). Multiply frequency by cost to get annual leverage. Anything above 200 hours a year is a strong first target.

Do not start with the most interesting problem. Start with the highest-leverage one. Interesting problems are what you graduate to.