Back to all of the ideas

Find your first automation

Score the tasks on your plate and see which one is worth handing to a robot first.

The first thing most people learn to cook is scrambled eggs. Nobody starts with a soufflé, and yet when people start with AI they tend to feel that the first project has to be impressive, like an agent or a complete end-to-end workflow.

That expectation keeps smart people stuck at the starting line. Scrambled eggs are simple enough to try today, forgiving enough that mistakes don’t matter, and useful enough that you’ll make them again tomorrow.

In my experience, teams rarely struggle with AI because they can’t write prompts. The harder problem is that they can’t confidently say what to automate or augment first. FAR is the filter I use to answer that, and it’s deliberately simple enough to run in your head during a hallway conversation.

  1. Frequent

    It happens daily or weekly, and the more often it comes around, the bigger the payoff.

  2. Annoying

    It is busywork, like formatting, chasing approvals, or retyping the same thing into different tools.

  3. Repeatable

    It has the same shape every time, and that consistency is what lets AI handle it reliably.

Score what is on your plate

Add a few tasks from your real week and rate each one from 1 to 5. If you get stuck, pick something that annoyed you in the last 48 hours. Your tasks are saved in this browser and are never sent anywhere.

  • Sample: Writing the weekly stakeholder update

    Sample

    This one is a sample that shows how the scoring works. Remove it once you have added your own.

    4
    4
    5

Your shortlist

  1. Start with this one

    Writing the weekly stakeholder update (sample)

    13 out of 15

    It fits a pattern that reliably ships: Draft, review, publish. AI writes the first version, a person reviews it, and then it goes out.

FAR finds the candidates, and impact picks the winner.

Check everything that would be true if your top task ran well.

Whatever you pick, choose the lightest shape that solves the problem. An assistant that helps you in the moment is quicker to try than a workflow, and a workflow is easier to trust than an agent.

Want to run this with your team?

I teach all of these as live sessions, where people work on their own real examples.

Bring this to your team