Before you automate, decide what deserves your attention.

Before choosing an AI tool, choose the piece of work you want to improve. What is difficult about it, and what would a useful result look like?

Start with one real task.

“Use more AI” is difficult to evaluate. “Turn an approved workshop outline into a first draft of a participant guide” gives you something specific to test.

Choose a recurring task with a clear input, an identifiable output and someone able to judge the result. Describe the current process first. Where do you spend time? Which parts involve repetition? Which parts need context, permission or a decision that remains yours?

Before automating a step, ask whether you still need to do it. An outdated report or unnecessary approval does not become more useful simply because it is faster to produce.

Decide what stays with you.

Write down the review points before you start. For a content task, these might include factual accuracy, whether the message fits the audience, whether the examples are yours to use, and whether the final text sounds like something you would say.

Make the boundary explicit: the tool can help prepare a draft. A person approves what leaves the business. The appropriate boundary depends on the task and its consequences.

For an initial experiment, use material you are comfortable testing with. Check the tool’s current data-handling settings and terms before providing confidential client information. An internal trial should not depend on using someone else’s private material.

Compare the whole process.

Try the workflow on a small, representative piece of work. Record the time spent preparing the input, checking the output and correcting it – not just the time the tool takes to respond.

Compare it with your usual method. Is the result useful? Did checking it take more effort than expected? Could someone repeat the process without relying on instructions that only exist in your head?

Keep what works, revise what does not, and decide whether another trial is worthwhile. A small process you understand is a useful starting point for deciding what to automate next.

Which recurring task could you test on a small scale – and what would make the experiment worth keeping?

Explore practical AI learning at the Academy, with a real task and your own judgement at the centre.