Learn Prompting 101 Iterating Your Way to Great Output

Iterating Your Way to Great Output

Beginner 🕐 15 min Lesson 7 of 10
What you'll learn
  • Treat every first output as a draft, not a finished product
  • Write effective follow-up prompts that are short and precise about what to change
  • Apply four types of iteration: refine, redirect, expand, and condense
  • Use alternatives prompts to generate options when you're not sure what you want
  • Know when to start fresh rather than continuing to iterate

The First Output Is a Draft

One of the most important mindset shifts in prompting is this: your first response is not the end of the process. It is the beginning of a conversation. Treating the AI's first output as a starting point — not a finished product — unlocks a whole new level of control over what you end up with.

Professional users of AI rarely write a single prompt and walk away satisfied. They write a prompt, review the output, identify what is missing or off, and send a follow-up. This process — called iteration — is often what separates people who consistently get great results from those who give up after one attempt.

What Makes a Good Follow-Up Prompt

Follow-up prompts are almost always shorter than the original. You do not need to restate everything — the AI remembers the full conversation. A good follow-up is precise about what needs to change:

  • "Make it shorter — cut to under 80 words."
  • "The tone is too formal. Rewrite it to sound more like a conversation."
  • "The third paragraph is weak. Rewrite just that section."
  • "Add a stronger call to action at the end."
  • "This is good but too generic. Make it more specific to our situation: we're a bootstrapped startup with three employees."

Each of these tells the AI exactly what changed. The AI applies that change while keeping everything else the same. That is far more efficient than rewriting your original prompt from scratch.

Types of Iteration

There are several distinct ways to iterate, each suited to a different problem:

Refine
Small adjustments

Output is mostly right. Target one specific element — a sentence, the headline, the opening.

Redirect
Change direction

Wrong angle, wrong tone, or it misread your intent. Steer it to the right track.

Expand
Add more

Output is good but incomplete. Ask for more depth, more examples, or a new section.

Condense
Remove the fluff

Output is right but too long. Cut to the key points without losing the ideas.

Refine: Small Adjustments

Use this when the output is mostly right but needs polish. A refinement prompt targets a specific element:

  • "Shorten the opening sentence."
  • "Replace the word 'utilise' with something simpler throughout."
  • "Make the headline punchier — give me three alternatives."

Redirect: Change Direction

Use this when the output went in the wrong direction — the tone is off, the angle is wrong, or it misunderstood your intent:

  • "This is too sales-heavy. I want it to feel educational, not promotional."
  • "You focused on features. Rewrite it focusing entirely on benefits."
  • "This reads like a corporate document. I want it to sound like a person talking."

Expand: Add More

Use this when the output is good but incomplete:

  • "Expand the second section — it needs more detail."
  • "Add a section on pricing considerations."
  • "Give me five more examples like the ones you listed."

Condense: Remove the Fluff

Use this when the output is right but too long:

  • "Cut this to half the length without losing the key points."
  • "Remove the bullet points and integrate the key ideas into two paragraphs."
  • "Keep only the three strongest points — cut the rest."

Asking for Alternatives

One of the most powerful follow-up prompts is simply asking for options:

"Give me five alternative versions of that headline, ranging from direct to playful."
"Rewrite that paragraph in three different tones: formal, conversational, and urgent."

Generating alternatives is something AI does extremely well and extremely fast. It is the equivalent of asking a copywriter for multiple options — except it takes seconds. Use this whenever you are not sure what you want until you see the choices.

When to Start Fresh

Iteration works best when the output is in the right ballpark. If it is fundamentally wrong — the wrong type of content, the wrong angle, a complete misunderstanding of the task — it is sometimes faster to abandon the conversation and start a new one with a better initial prompt.

Signs you should start fresh:

  • You have sent more than four follow-up prompts and the output is still wrong
  • The AI seems to be "stuck" on a framing you gave it early in the conversation
  • The output type is wrong (e.g., you wanted a formal report and got a casual blog post)

When starting fresh, use what you learned from the failed conversation. Ask yourself: what was missing from my original prompt that caused it to go wrong? Add that to the new version.

Saving Your Best Prompts

When you find a prompt + iteration sequence that works well, write it down. Your best prompts are reusable assets. A prompt that produced exactly the right type of email for your business will work again next time you need that email. Building a small personal prompt library — even just a notes document — multiplies your productivity over time. Lesson 10 covers this in more detail.

Key takeaways
  • First output = draft. Iteration is how you get to the final version
  • Follow-up prompts are shorter than the original — the AI remembers the conversation
  • Four modes: refine (small fixes), redirect (wrong angle), expand (add more), condense (cut fluff)
  • 'Give me five alternatives' is one of the most powerful follow-up prompts
  • If still wrong after four iterations, start fresh with a better initial prompt