Best Claude prompts for makers and PMs

Here is a one-line brief for a feature: [brief].

My product: [one sentence on what it does and who it's for].
Our biggest constraint: [engineering bandwidth / regulatory / performance / etc.].

Write a PRD that includes: problem statement, success metrics (3 max), proposed solution, out of scope, open questions. Format as a structured document I can share with engineers.

Flag any assumption you're making about user behavior or technical feasibility.

Here is the proposed solution section from a PRD: [paste].

Write user stories in "As a [persona], I want [action], so that [outcome]" format. Each story needs at least two acceptance criteria written as testable "Given / When / Then" conditions. Persona options: [list your actual personas].

Reject any story that would require a separate PRD to spec properly — break those out as a list at the end.

Here is a PRD for [feature name]: [paste].

Write a list of questions an engineering lead would ask before committing to a timeline. Organize by category: data model, third-party dependencies, edge cases, performance, rollback. No answers — questions only.

Here is a PRD draft: [paste].

Play the role of a skeptical senior PM who has shipped this category of product before. Give me your three biggest concerns — not general PRD quality issues, but specific gaps given what I've written. Quote the section you're responding to.

Here are five user interview transcripts: [paste].

Synthesize them into: (1) three to five distinct pain points with direct quotes supporting each, (2) one insight that surprised you — something that didn't match a conventional assumption about this user, (3) the one question these interviews left open.

Do not make up quotes. If a theme appears in fewer than two interviews, say so.

Here are 40 open-text responses to the question "[your survey question]": [paste].

Cluster the responses into no more than six themes. For each theme: a label (3 words max), a representative verbatim quote, and an approximate percentage of responses. List any responses that didn't fit a cluster.

Here is a summary of our user research findings: [paste].

Write two to three user personas. For each: a name and role, their primary goal when using [your product], their biggest frustration with current tools, and a verbatim-style quote that captures their mindset. Base these strictly on the research provided — don't add archetypal PM/designer personas unless the research mentions them.

We're running discovery interviews to understand [problem area] among [user type].

Write a 45-minute interview guide. Include: an opener that builds rapport and explains the session is about their experience, not a product test; five to seven core questions that elicit stories rather than opinions; two probes per question to go deeper. Avoid any question that could be answered yes or no.

Here is a structured comparison of four competitors: [paste your notes or feature matrix].

Identify: (1) the one axis where we are genuinely ahead, (2) the one axis where a named competitor is genuinely ahead of us and customers care about it, (3) the whitespace — a segment or use case none of them serve well. Be honest about #2; I need this for internal strategy, not sales copy.

Our product: [one sentence].
Our three strongest competitors: [names + one sentence each].
What we do better: [specific].
What we lose on: [honest].

Write three positioning statement options, each targeting a different buyer emotion: one for the person burned by the incumbent, one for the person evaluating rationally, one for the person who needs to justify the switch to a team. Each should be under 30 words.

Here is a raw set of facts about our product and market: [paste].

Write a SWOT analysis. Each quadrant: four items, ranked by importance. For every Weakness, write one mitigation we could realistically attempt. Do not include generic items like "strong team" or "growing market" unless the evidence above supports it specifically.

Here is a [PRD / research report / strategy doc]: [paste].

Write a four-paragraph executive summary for a VP who will spend two minutes reading this. Paragraph one: what we decided and why. Paragraph two: what we're not doing and why not. Paragraph three: what we need from them. Paragraph four: risks and how we're handling them.

Here is a raw brain dump of where the [project name] project stands this week: [paste notes].

Rewrite this as a status update for a weekly all-hands. Format: one sentence summary (green / yellow / red), three bullet points on what happened, one bullet on the main risk, one bullet on what we need from other teams. Total under 150 words.

I need to present [topic] to [audience: board / investors / eng team / cross-functional].

The decision I want them to make by the end: [specific decision].

Write a ten-slide narrative outline — not bullet points on each slide, but the *story* the deck tells, one sentence per slide. Then write the three strongest objections this audience is likely to raise and a one-sentence response to each.

I need to tell [stakeholder role] that [the bad news].

Context: [what happened, what led to it, what we've already tried].

Write a message that: leads with the impact, not the backstory; owns what went wrong without hedging; states clearly what happens next. Under 200 words. Don't soften the central fact.

We are launching [feature name] on [date]. It does [one sentence on what it does]. The users who will care most: [segment]. The risk of the launch: [specific].

Write an internal launch email for the company. Tone: direct, not celebratory. Include: what it is, who it's for, what to watch in the first week, and the one thing we're most worried about.

Feature: [name and what it does].
Who it's for: [user type].
The problem it solves: [specific].
One concrete thing users can do now that they couldn't before: [specific].

Write a changelog entry under 100 words. No adjectives that don't add information. No "we're excited to announce."

We are launching [feature] to [audience] on [date]. Our product is [description]. Our stack involves: [relevant tech].

Write a launch checklist organized into: T-7 days, T-2 days, day of, T+3 days. Include the categories most relevant to a product launch in a [B2B SaaS / consumer / developer tool / etc.] context. Flag any item that is easy to forget but high-consequence.

Product: [name and one sentence on what it does].
Primary user: [who, their role/situation].
The main alternative they use today: [competitor or workaround].
The thing we do that they can't: [specific].

Write a landing page hero section: headline (under 10 words), subhead (under 25 words), three value proposition bullets (each under 12 words). Avoid "streamline," "seamless," "powerful," and any phrase that would also apply to a spreadsheet.

I'm reaching out to [recipient type] about [product]. They don't know me.

The one thing I want them to do after reading this: [specific action].
What's in it for them, specifically: [concrete benefit, not a feature list].

Write a cold email under 80 words. Lead with something relevant to them, not to me. No "I hope this finds you well." No product feature list. One CTA.

Our pricing is: [tiers, prices, what's in each].
The most common objection at checkout: [what you've heard].
The user most likely to upgrade from free to paid: [describe them].

Write copy for a pricing page: a one-line pitch above the tiers, two to three words that label each tier (not "Basic / Pro / Enterprise"), and one sentence of social proof or risk-reduction copy below the CTA.

A user said: "[verbatim feedback or cancellation reason]."

Write a reply that: acknowledges their specific reason (not generic), offers one concrete thing we're doing or planning to address it if true, and keeps the door open without being needy. Under 100 words.

Here is a draft: [paste].

Rewrite it at 70% of the word count. Prioritize cutting: throat-clearing openers, phrases that restate what was just said, any sentence that wouldn't be missed. Do not cut any specific fact, number, or decision. Show me the cuts you made as a diff if possible.

Here is an argument I'm making to [audience]: [paste].

Identify: (1) the one sentence that most undermines my credibility — the place where I'm overstating, hedging badly, or making a claim I haven't supported; (2) the counterargument I haven't addressed; (3) the sentence I could cut that would actually strengthen the point.

Here is a technical explanation of [thing]: [paste].

Rewrite it for a non-technical reader who has never heard of [technical term]. No jargon, no acronyms without expansion. Target: a smart 16-year-old who cares about the problem but not the implementation. Under 150 words.

You are my PM assistant for [product name].

Product: [2–3 sentences on what it does, who it's for, and the core value proposition].
Users: [describe the primary user segment and their main job to be done].
Competitors: [list 2–3, with one sentence on what each is known for].
Our positioning: [where we win, where we concede].
My role: [your role and what you typically ask Claude to help with].
Tone: [direct, no hedging, flag when you're uncertain].

When I paste raw material (transcripts, data, notes), default to synthesis — pull out the signal and surface what's non-obvious. When I ask you to write, default to the format I describe or ask me before writing prose.

Here's what you wrote: [paste Claude's output].
Here's what I actually need it to do differently: [specific].
Here's why the previous version missed: [specific].

Revise it with those constraints. Don't re-explain what you changed — just show me the new version.

I want a working prototype of [feature]. It should: [list 3–4 core interactions].
Use: HTML, CSS, and JavaScript in a single self-contained file. No external dependencies except [CDN library if needed].
Do not build: [what to leave out to keep it scoped].
The prototype is for [stakeholder type] who will [what they'll do with it]. Optimize for clarity of the core interaction, not production quality.

Here is a [language] function that [what it does]: [paste code].

Review for: (1) the most likely runtime bug; (2) the edge case I probably haven't handled; (3) one thing that will make this hard to maintain in six months. Give me the line number and a one-sentence fix for each.

I'm about to [decision / action / launch / write]. Here is everything I know about the situation: [paste].

What am I not considering? Be specific — not general advice, but what is missing or wrong given exactly what I've described. Rank the gaps by how much they could change my decision.