Before Automating, Fix the Data Underneath

A manager wants a reminder to go out automatically. A report should update itself. A document should route to the right place. A lead should move from enquiry to follow-up without being copied twice.

All of that sounds like automation.

Then you look at the data. The same client has three spellings. Dates are typed into notes. Status is implied by colour. The important file is attached to an email. The spreadsheet has columns nobody trusts.

Automation is just a set of rules. If this happens, do that. If the record has this value, move it there. If the deadline is tomorrow, send the reminder.

That only works when the signal is clear.

If the deadline is hidden in a PDF, if the status lives in someone's memory, or if the same field means three different things, automation becomes fragile. Someone will end up checking the automation manually, which defeats the point.

This is where many small businesses get stuck. They think the data problem means a database problem.

Sometimes it does. But often the data is the messy operating material around the work: emails, PDFs, booking exports, invoices, WhatsApp messages, folder names, report tabs, audio notes, scanned documents, and old spreadsheets.

The job is not to make everything fancy. The job is to make the important information usable.

Before automating, agree the basic structure of the work.

You need enough clarity that a system can tell what record it is looking at, what state it is in, what source supports it, and what should happen next.

Structure the records first. Fix the workflow next. Add automation after that.

That order is slower at the beginning and faster everywhere else. It also gives you a system that can be supported, explained, and improved later.