Nudge Forward

In the olden days, like last year, I’d learned that AI requires careful prompting.

Giving it a persona and detailed directions.

As a result, my standard guidance when asking AI to proofread was:

You are an expert and very pedantic proofreader. Very carefully read the following update. Then accurately list every possible spelling mistake as well as any possible grammatical errors or suggested improvements. Summarise your recommendations in a table format making it clear, quick and easy to review. The text is as follows: [text]

What a difference a year makes.

These days life’s a lot simpler. As the tools get smarter and I become wiser.

This year I’m doing less prompting, more nudging. Streamlining my proofreading prompt into:

Proofread: [text]
List changes

Discovering that the follow-up second line is the critical one.

As the AI models and applications evolve, some now clearly detail the editing changes made without prompting. Which is helpful but sometimes more time‑consuming than needed. So other AIs now skip this step by design – just as some people provide a rationale without being asked, while others don’t.

My immediate reaction when I encountered this variation between AIs was to think that one was now better than the others.

Until I realised that as the technology evolves, so must we.

Asking better questions. Adapting our leadership style depending on who or what we’re working with.

Less detailed prompting. More thoughtful nudging.

Forward.