5 simple steps to improving your AI prompting

We’ve learnt to get better outcomes from AI by improving our prompting...

If you’ve ever typed a question into ChatGPT and thought “urgh, that’s not what I wanted”, you’re not alone.

That’s where RCOTE comes in: a simple 5-step prompting framework we use on our course.

1. Role – give the AI a role. (Who should they be?)
2. Context – what’s the situation or background the AI needs to know?
3. Objective – what outcome do you want?
4. Tasks / Instructions – what specific steps should it follow?
5. Expectations & Format – how should the answer look, what’s important?

Example prompt using RCOTE…

ROLE: You are a world-class career coach in sustainability.

CONTEXT: I’m a mid-level project manager in a renewable energy startup exploring leadership roles.

OBJECTIVE: I want to map out my next career steps.

TASKS: Please suggest 3 pathways, each with skills I should build and common pitfalls to avoid.

EXPECTATIONS: Present it in a table format, concise but actionable.

Result? Instead of generic waffle or made up rubbish, you’ll get structured, relevant advice you can act on.

Less hallucinations, less AI processing (energy use), better results.

This is the kind of practical skill we teach (and practice together) inside our Good with AI cohorts.

Cohort 4 + 5 start in just over 2 weeks - join us if you want to turn AI into a helpful teammate 👉 https://goodwithai.org/