How to spot the tasks actually worth automating, in ten short questions — five for you, five for your team. About three minutes. Biscuit optional.
Most automation projects flop for the same boring reason. People automate the wrong task.
They pick the task that sounded impressive at a leadership offsite. Or the one their nephew’s friend automated at a startup nobody can remember the name of. Or - most tragically - the one that was already taking five minutes a week and after a six-month project, it now takes… four minutes a week. Oh dear!
The tasks actually worth automating are quieter. They hide in your calendar, in your team’s polite sighs,
and in the little spreadsheet someone has been maintaining since 2019 because “that’s just how we do it.”
Here’s how to find them, without a discovery workshop, a consultant in a turtleneck, or a single Post-it note.
Ask yourself (as the manager)
You’re closer to the bottleneck than you think. Before you go interview anyone, sit with these:
- If a magic wand could vanish one recurring task from your week, which one goes? The one you pictured first. Not the second one. The first one.
- Which task do you redo every week with only the names and numbers changed? Anything that rhymes with last week is a candidate. Anything that rhymes with last month is a layup.
- What are you doing instead of the work you were actually hired for? If you were hired to grow the business and Tuesday vanished into formatting a report, write that down.
- Which task makes you procrastinate hardest? Your dread is data. The thing you keep “getting to after lunch” is usually mechanical, repetitive, and ripe.
- What would you love to never explain to a new hire again? If onboarding someone into a task feels worse than doing it yourself, the task is over-engineered for humans.
Ask your team
This is where the real gold lives. Your team knows exactly which tasks are slowly turning their brain to soup - they’re just waiting to be asked properly. Try these in a 1:1, not a group meeting (people are more honest (and funnier) one-on-one):
- What’s the task you could do blindfolded, hungover, on a moving train? Steady-state, repetitive, no judgment calls. This is the gold standard.
- What’s the most boring 30 minutes of your week? A specific number forces a specific answer. “All of it” doesn’t count, though you should probably also have a chat about that.
- Which task makes you mutter “there has got to be a better way”? If they’ve said it out loud more than three times, it’s been a problem for a year.
- If we hired a very enthusiastic intern tomorrow, what would you hand them on day one? Whatever they’d hand off is something you can probably hand off to an AI instead. The intern will thank you. The AI will not, but the AI is fine.
- What’s the task you’d happily teach a robot - even if it took a whole afternoon - if it meant never doing it again? The willingness to invest the afternoon is the signal.
What to do with the answers
You’ll notice patterns. The same three tasks will show up across half your team. The thing you put first on your list will probably be the thing your most senior person put first on theirs too. That’s not a coincidence - it’s a queue.
Start with the task that:
-
shows up on more than one list,
-
has a clear input and a clear output (data goes in, deliverable comes out),
-
and doesn’t require reading a customer’s mood or making a judgment call that could end up in a lawsuit.
Save the judgment calls for the humans. Hand the copy-paste to the machines. Repeat until your calendar stops giving you the ick.
Want help turning your list into actual working automations without spending three months on a discovery phase?
That’s basically our favourite conversation to have. Book a call - we keep the tech jargon to a minimum and the existential dread to zero.