Automation isn’t about replacing people, it’s about freeing them from the grind.The challenge is knowing what’s actually worth automating, because not everything should be.
Here’s a short, practical guide to help you spot what can (and should) be automated or digitised – and how to work out if it’s worth the effort.
1) Map Your Week – Find the Repeatables
Start by mapping out your typical week. Write down every task you or your team repeat more than a couple of times a month. You’ll be surprised how much time hides in the routine.
Examples:
- Filling drawings, renaming, and re-saving versions
- Writing site reports or progress summaries
- Copy-pasting data between Excel sheets or emails
- Collecting signatures, printing, scanning, re-uploading
- Updating trackers or QA logs
Each of these tasks eats 30-60 minutes, but happens dozens of times a month. That’s where automation earns its keep. If you’re after more ideas for what you could automate, try here: Inspire
Don’t start with design tasks. They need creativity and human judgement. Start with admin work, and communication loops – the repeatable stuff that quietly steals hours. Automating high-judgement engineering or design work carries risk, while many archaic processes still have huge automation or digitisation ROI.
2) Score It – Quick and Objective
Once you’ve got your list, give each recurring task a quick score. This helps you prioritise what’s worth tackling first.
| Task | Time Spent (1–5 higher for more time) |
Frequency (1–5 higher for frequent) |
Repeatability (1–5 higher if identical every time) |
Risk (1–5 higher for low risk) |
Score (multiply the scores) |
Priority |
|---|---|---|---|---|---|---|
| Site inspection report | 5 | 4 | 4 | 2 | 160 | High |
| Drawing revisions log | 3 | 5 | 5 | 2 | 150 | Medium–High |
| Complex design calculation | 2 | 1 | 1 | 5 | 10 | Skip |
| … | ||||||
| … |
Estimate the Payback
Use this quick logic to check if an automation is worth it:
Cost saved = (hours saved × times per month × hourly rate) – setup time
If the payback period is less than 6 months, it’s worth doing.
Example: a 45-minute report now takes 5 minutes, done 30 times a month by a £40/hr engineer, that’s roughly £800 saved per month, every month.
The Takeaway
Don’t automate everything. Automate what’s frequent, boring, and measurable.
Start small. Measure the win. Repeat what works.
That’s how real digital transformation happens, one smart system at a time.
For clarity on what I do once I know what I want to automate, have a look here.
P.S. Where AI Fits In
AI works best when it supports structured automation, not when it replaces it. Think of it in two layers:
- Rules-based automation: file naming, dashboards, reminders, data transfers.
- AI-assisted automation: drafting reports, summarising site notes, generating proposal text.
Start with the rule-based foundation. Once it’s stable, layer AI on top to make it smarter.