Prompting as a Professional Skill · Prompt Engineering

Clear thinking, expressed clearly.

Copilot reveals where your thinking is fuzzy, hard-to-specify prompts flag hard-to-think-clearly problems.

Prompting as professional skill

Clear thinking expressed clearly, deliberate reps on real work.

Core principles

Fluency looks like: diagnosing weak outputs quickly, reaching for the right framework without overthinking, refining in two or three moves, maintaining a growing library.
Deliberate practice: one technique per week on real work; monthly review of library; teach one pattern to a colleague.
Professional standard: you own what you send, prompting accelerates drafting; judgment and verification stay yours.
Next steps: deepen with AI Judgment for trust; build agents when a prompt should be reusable for the team.

Go deeper: Getting Started: prompting

Check yourself

What does this lesson say hard-to-specify prompts usually signal?

Do this in Copilot

Choose one workflow you will run only through your library for two weeks. Note time saved and quality vs starting from scratch.

Paste this into Copilot Chat and work through it before moving on.

Weekly practice commitment

For the next two weeks I will use my prompt library for [SPECIFIC WORKFLOW, e.g. stakeholder emails, meeting recaps]. After each use, note: did the saved prompt need refinement, and what one line would improve the library entry?

Open Copilot →

Deliberate practice log

For the next two weeks I will use my prompt library for [SPECIFIC WORKFLOW, e.g. stakeholder emails, meeting recaps]. After each use, note: did the saved prompt need refinement, and what one line would improve the library entry?

Did you run this in Copilot? Mark complete when you have tried it.

The longer game

You just learned that vague inputs produce vague outputs, and that specificity, goal, context, expectation, closes that gap. That works. Use it.

Here's what it points toward. The same research that found 90% of large companies underperforming on AI readiness identified the root cause: not access, not training completion, not tool sophistication. The gap was whether people had genuine freedom to experiment, to try something before it was perfect, to test an idea before they knew how it would land, to iterate in public rather than wait until the output was polished. That's not a prompting skill. It's a disposition. The frameworks you just learned are the mechanical part. The harder part is being the person who actually uses them on the work that matters, not just the safe tasks, not just when the output is already mostly there, but early, imperfectly, on the thing you actually care about getting right.