Course Home Weak vs. Strong Prompt Examples

Weak vs. Strong Prompt Examples

A practical comparison guide

This document shows the same request written as a weak prompt and a strong prompt. The purpose is not to show extreme opposites — the weak versions here are the kind of thing a thoughtful person might actually write. The strong versions show what adding audience, context, goal, and constraints actually looks like.


Example 1: Blog post

Weak:

Write a blog post about customer onboarding.

Strong:

Write a 700-word blog post for a B2B SaaS company. Audience: customer success managers who work at software companies. Topic: reducing churn during the first 30 days of onboarding. Tone: direct and practical — no hype, no fluff. Assume the reader has tried basic onboarding tips before and wants more specific, actionable ideas. Avoid generic advice like "check in frequently." Give concrete examples where possible.

What changed: Audience named. Goal defined (reduce churn). Tone set. Generic advice explicitly excluded. Expertise level assumed.


Example 2: Email

Weak:

Write an announcement email about our new feature.

Strong:

Write a short announcement email for existing customers about a new bulk export feature we just launched. These customers have been asking for this feature for months — it lets them download multiple reports at once. Audience: power users who are familiar with the product. Tone: direct and friendly — not hypey. Under 150 words. Subject line should be specific and clear, not promotional. Avoid phrases like "exciting news" or "we are thrilled."

What changed: Audience narrowed. Feature described. Tone and style guided. Length capped. Specific phrases excluded.


Example 3: Meeting summary

Weak:

Summarize these meeting notes.

Strong:

Summarize these meeting notes into a brief internal update for team members who did not attend. Focus on: decisions made, action items with owners, and open questions that still need resolution. Format: three labeled sections. Tone: direct and factual — no narrative. Under 250 words. [Notes pasted below]

What changed: Audience defined (non-attendees). Specific capture goals named. Format structured. Length limited.


Example 4: Job description

Weak:

Write a job description for a marketing manager.

Strong:

Write a job description for a senior content marketing manager at a B2B software company with a remote-first team of 25 people. We want someone with 5+ years of experience who can own content strategy independently. The tone should be direct and reflect our no-BS culture — skip inflated corporate language. Include sections for: role overview, key responsibilities (5 bullets max), required qualifications, and why this role is interesting. Avoid language that implies we want someone junior. Under 500 words.

What changed: Company context added. Experience level specified. Culture guidance included. Sections defined. Anti-junior signal given. Length capped.


Example 5: Research summary

Weak:

Summarize this research on remote work productivity.

Strong:

Summarize this research on remote work productivity for the leadership team at a 200-person software company. They are deciding whether to introduce an optional return-to-office policy. Focus only on findings that speak to productivity and collaboration — skip methodology, academic framing, and anything about employee wellness (covered separately). Format: three to five clear findings as bullet points, each with a one-sentence "implication for our decision." Under 400 words. [Research pasted below]

What changed: Audience defined. Decision context given. Focus area narrowed. Exclusions named. Actionable framing added. Length set.


Example 6: Project update

Weak:

Write an update for the stakeholders about the project.

Strong:

Write a brief stakeholder update about a software release that has been delayed by two weeks due to QA findings. Audience: internal leadership team. They are supportive but want clear facts and the new timeline. Tone: candid and confident — not apologetic or defensive. Format: three short paragraphs — what happened, what we found, what the updated plan is. Under 200 words. Avoid hedging language like "it seems" or "we believe we can."

What changed: Situation explained. Audience understanding noted. Tone characterized in positive and negative terms. Structure specified. Hedging explicitly excluded.


Example 7: Explanation for a non-technical audience

Weak:

Explain what an API is.

Strong:

Explain what an API is to a non-technical marketing manager who needs to understand it well enough to participate in a meeting with an engineering team. Focus on what APIs do at a practical level — not how they work technically. Use an analogy if it helps. No jargon. Two short paragraphs at most.

What changed: Specific audience role given. Purpose of the explanation named. Level of depth defined. Technical depth explicitly excluded. Length capped.


Example 8: Strategy memo

Weak:

Write a memo recommending we move to a tiered support model.

Strong:

Write a one-page decision memo recommending we shift from a generalist support model to a tiered specialization model. Audience: founding team who will approve or reject the change at the next all-hands. They care about customer impact, operational feasibility, and whether the timing is right. Structure: current state, key limitation, recommendation, why now, and the specific decision being asked for. Tone: direct and confident — no hedging. Under 400 words.

What changed: Decision context built in. Audience decision criteria named. Structure prescribed. Action ask required. Hedging excluded.


Pattern summary

What was added Why it matters
Specific audience role Changes vocabulary, depth, tone, and framing
Clear goal or outcome Focuses the document on a purpose, not just a format
Tone description (what it should/should not feel like) Prevents generic corporate writing
Format instruction Avoids AI inventing structure that does not fit the situation
Length limit Forces prioritization, prevents padding
Exclusion instructions Eliminates the most common unwanted AI defaults
Context (project, situation, decision) Grounds the output in something real instead of generic
Working Well With AI · Practical AI training for real work