66. Iterative Prompt Improvement

The lesson

Iteration is key: Your first prompt rarely gives perfect results. Professional prompt engineering is iterative refinement.

Step 1 - Analyze the response: What was good? What was missing? Was the format wrong? Did it make assumptions?

Step 2 - Identify the gap: Common issues: missing context, ambiguous instructions, wrong persona, unspecified format, or too broad a request.

Step 3 - Make one targeted change: Do not change everything at once. Fix the most important issue first, then test again.

Example iteration: First attempt: 'Write about dogs' → Result: Too generic. Second attempt: 'Write a 200-word introduction to dog training for first-time owners' → Better but missing structure. Third attempt: Add 'Format as 3 bullet points: key preparation, first week essentials, common mistakes to avoid' → Targeted and useful.

Keep a log: Note what changes improved the output. This builds your personal prompt engineering intuition.

A junior data analyst refines a report prompt over three rounds in TPEE, fixing one issue at a time and noting which wording produced the clearer table.

A restaurant chef iterates a menu-description prompt, first adding a word limit, then a short bullet structure, and keeps the version staff found easiest to read.

A claims assessor improves a summary prompt by adding a missing audience line, tests it again, and records the change in a simple log for the team.

A laboratory technician revises a method-summary prompt after seeing the first answer, adding an explicit output format before running it a second time.

A news reporter iterates a headline-generation prompt, tightening the instruction after each weak result and keeping the clearest version for later stories.

A warehouse shift coordinator keeps a running log of prompt changes, noting which edit to a stock-report prompt produced a more useful list.

A physiotherapist refines a home-exercise prompt by making the audience clearer, testing the change before saving the improved version for patient handouts.

A primary school teacher improves a lesson-summary prompt by adding two short examples, then checks whether the next answer matched the intended structure.

A town planner iterates a public-notice prompt, changing only the format line each round until the draft reads clearly for residents.

The owner of a small bakery refines a specials-board prompt over several tries, adjusting one detail at a time and saving the wording that worked best.

Check yourself