Automate the Boring Bits First

Somewhere in your business there's a task that takes forty minutes a week and involves copying information from one place into another. Someone good at their job does it carefully, every Friday, and has done for two years.

That task is the best automation candidate in the business. It's also the one nobody proposes, because it isn't interesting enough to bring up in a meeting.

Which is backwards. The most valuable automation is almost always the most boring one, because boring work is the work nobody defends when it goes.

Automation rewards clarity and punishes ambiguity

Here's the thing most people get wrong about automation: it doesn't make a bad process good. It makes whatever you've got happen faster and more often.

If your process is clear, automation removes the labour and the error rate. If your process is fuzzy - if it depends on judgement, exceptions or who's doing it - automation surfaces every inconsistency at speed, and you'll spend longer fixing it than the task ever took.

So the first question is never "what should we automate?" It's "is this clear enough to automate?"

The three tests

A task is ready to automate when it passes all three:

• It runs on a trigger - a schedule, a form submission, a status change, an email arriving. Something identifiable happens first
• The inputs are structured - the information it needs exists in fields, not buried in a paragraph or someone's memory
• The rule is unambiguous - a competent person would reach the same outcome every time, without a judgement call

Fail any one of those and you have a process problem, not an automation opportunity. Fixing it is usually the better job anyway: a process clear enough to automate is also clear enough to hand to someone else.

Where to look first

• Copying between systems - if a person is transferring data by hand, something is listening that could do it
• Recurring reminders and checks - weekly reports, follow-ups, review dates, renewals. Anything relying on someone remembering
• Status updates and notifications - telling a client their job is scheduled, telling a team a task is ready
• Document generation - quotes, invoices, reports assembled from data you already hold
• Onboarding sequences - new client or new starter steps that are identical every time and currently run from a list in someone's head

Notice what isn't on the list: anything requiring judgement about a client, a price or a person. Those stay human, deliberately.

How to automate without breaking things

This runs through the Implementation & Integration and Sustained Performance phases of our Operational Excellence Framework.

1. 1

   Fix the process first. Map it properly and remove anything that's only there through habit. If it isn't clear on paper, it won't be clear in an automation.

2. 2

   Start with one workflow. Pick the task that's most frequent and least ambiguous. Get it working end to end before touching anything else.

3. 3

   Build where your team can see it. Use integrations and automations inside tools your team already understands, rather than something only one consultant can maintain. Make, Zapier and the automation features in ClickUp and Monday.com all cover this.

4. 4

   Keep a human check on the output. For the first few weeks, someone reviews what the automation produced. You're looking for the cases it didn't expect.

5. 5

   Write down what it does. One page: what triggers it, what it does, what happens if it fails, and who owns it. Six months from now that page is the only reason it still works.

6. 6

   Measure the time back. Know what it used to cost you. It's the only way to know whether the next automation is worth doing.

The bit most people skip

Automation changes what people do, and that needs handling properly.

If someone's been doing a task for two years and it disappears into a system with no conversation, they'll reasonably wonder what else is being taken off them. The businesses that get this right treat it as capability building: the tedious part goes, and the person moves to work that uses more of their judgement.

That's the actual return. Not the forty minutes. The forty minutes every week, plus a person doing work worth paying them for.

What this looks like in practice

A professional services client of ours was producing a monthly client report by hand - pulling figures from three places, formatting them into a template, checking them, sending. It took most of a day and it happened twelve times a year, plus whenever a client asked early.

The figures were already structured and the format never changed. So the process was cleaned up first - one source for each figure, one definition of each measure - and then assembly and sending were automated. The checking stayed with a person, deliberately. The day came back, and the report stopped being the job everyone quietly dreaded.

Quick self-check

If you answer "no" to more than one, you're probably automating around unclear processes - which costs more time than it saves.

Your task for this week

Find the task someone does every week that involves copying, retyping or reformatting. Write down what triggers it and what a correct result looks like. If you can fill in both without hesitating, it's ready. If you can't, you've found the process to fix first.