Which tasks can't I automate?

Work that genuinely needs human judgement, creativity or a personal connection. And tasks that change so often the automation can't keep up. The rest is often surprisingly easy to automate.

What if a task can only be half automated?

That's fine, and even common. Often you take away 80% of the work and the human check remains. That 80% is the gain.

How many tasks should I tackle at once?

One. Finish it, show the result, and use that momentum for the next. Ten at once rarely leads to anything that keeps running.

Guide · 1 min read

Automating tasks: where do you start?

Not everything at once, but the right task first. How to spot which repetitive jobs lend themselves to automation, and which you're better off leaving alone.

The biggest mistake in automating is wanting to start with everything at once. The gain is in choosing: one task, done well, that saves time every week. Then the next. This page helps you find the right first task.

The three traits of a good candidate

A task lends itself to automation if it is:

Repeatable: you do it often and in the same way
Predictable: there are few erratic exceptions to it
Time-consuming or error-prone: there's something real to gain

Tick all three and you've got a winner.

Make a short list

Walk through a normal working week and note down everything that feels like "I do this every single time". Retyping, forwarding, looking up, reporting, reminding. Estimate per task: how much time does this cost per week? How annoying is it? How often does it go wrong?

Choose on gain, not on difficulty

Plot the tasks against each other on two axes: how much time it saves, and how much effort it takes to automate. Start in the corner with high gain and low effort. That delivers a tangible result quickly, and that result convinces your team more than any plan could.

For inspiration: take a look at the examples of AI automation. A common starting point with a lot to gain is automating email.

What you're better off leaving alone

Not everything belongs automated. Work that genuinely calls for a human, a tricky client conversation, a creative trade-off, a sensitive decision, you leave be. And a task that's different every month costs more in upkeep than it delivers. Being honest about what isn't worth it is part of the approach.

From list to working automation

Have you chosen your first task? Then you can build it yourself with no-code automation, or hand it over. We help choose and build, and if a task isn't worth it, we'll say so.