Asset 20 8 2
Does AI recommend your business? Run the free check →

Join 15,000 business owners, marketers and entrepreneurs. The Sunday newsletter you'll be annoyed only arrives once a week.

Article

Claude Prompts for Product Managers That Work

If you are skim reading
The short version: Claude is better than most other AI tools at long, messy product documents because it can hold an entire PRD, research transcript or roadmap in context without losing the thread, so the prompts that work best for product managers lean into t

The short version: Claude is better than most other AI tools at long, messy product documents because it can hold an entire PRD, research transcript or roadmap in context without losing the thread, so the prompts that work best for product managers lean into that strength rather than treating it like a search engine. Below are prompts built for the actual work product managers do every week, from writing specs to prioritising backlogs to translating engineering jargon for stakeholders.

Why Claude specifically, for product work

I get asked a lot why I point product managers towards Claude rather than ChatGPT. The honest reason is context length and tone. Claude 3.5 and the newer Claude models handle long pasted documents, like a 4,000 word customer interview transcript or a sprawling Jira export, without flattening everything into generic bullet points. ChatGPT is still excellent for product work too, and if you want a comparison I covered separately, AI prompts for product managers using ChatGPT is worth reading alongside this one. But for the specific job of reading a pile of raw material and producing a structured, human-sounding document out the other end, Claude tends to win.

The uncomfortable bit nobody tells you: most product managers use Claude badly because they ask it to make decisions it has no business making. "Should we build feature A or feature B" is not a prompt, it's an abdication. Claude does not know your churn numbers, your sales pipeline, or that your CEO promised feature A to a big account last quarter. The prompts that help are the ones where you feed Claude the facts and ask it to organise, structure, or stress-test your thinking, not the ones where you ask it to think for you.

Writing a PRD from scratch, worked example

Say you run a mid-size SaaS product, a project management tool with around 40,000 active users, and your team has decided to build a "client view" where external clients can see task progress without logging into the full app. You have a messy page of notes from a stakeholder call. Here's the prompt:

  • "Here are my raw notes from a stakeholder call about a new client-facing view feature [paste notes]. Write a one-page PRD with these sections: Problem statement, Target user, Success metrics, Scope (in and out), Open questions. Keep the problem statement to three sentences max. Flag anywhere my notes contradict themselves."

That last instruction matters more than people realise. In a real version of this exercise, notes from different stakeholders often contradict each other, one person wants clients to comment on tasks, another explicitly says "view only, no interaction." Claude will catch that if you ask it to, and it will miss it completely if you don't. That single sentence, asking it to flag contradictions, is the difference between a PRD that saves you a meeting and one that causes a worse one three weeks later.

Turning a backlog into a prioritised shortlist

Paste your entire backlog, even if it is 60 items long and formatted badly, and use this:

  • "Here is my product backlog [paste]. Group these into themes. For each theme, estimate rough effort as Small, Medium or Large based on the description, and note which ones seem to depend on another item being done first. Do not suggest a priority order yet, just organise."

Then in a second message: "Now, using RICE scoring (Reach, Impact, Confidence, Effort), score the top three items in each theme and show your working." Doing it in two steps rather than one stops Claude from rushing to a tidy-looking answer before it has understood the shape of the backlog. I've found two-step prompts consistently outperform one long one for anything involving judgement calls.

Stakeholder updates that don't sound like a press release

This is the one most product managers get wrong. They ask for a "status update" and get back something that reads like a LinkedIn post. Try this instead:

  • "Write a stakeholder update for this sprint [paste your raw sprint notes, including anything that went wrong]. Audience is a VP who skims. Lead with the one thing that changed since last update, not a summary of everything. Maximum 150 words. No phrases like 'exciting progress' or 'great momentum'."

Banning the fluff phrases outright in the prompt works far better than editing them out afterwards, because Claude, left unchecked, defaults to corporate cheerfulness the same way most AI tools do.

A step by step for running a feature spec review

  1. Paste the draft spec and ask Claude to list every assumption it can find that isn't backed by data in the document.
  2. Ask it to write three questions an engineer would ask before estimating this.
  3. Ask it to write one question a skeptical executive would ask.
  4. Paste the answers back in and ask for a revised spec, tracking what changed.
  5. Read the final version yourself. Do not skip this step. Claude is a very good first draft machine and an unreliable final check.

Translating engineering speak for non-technical stakeholders

If part of your job involves explaining technical debt or architecture decisions to people who sit in finance or sales, this prompt earns its keep:

  • "Explain this engineering update to a non-technical sales director [paste the ticket or Slack thread]. Use one analogy, not more. Say what this means for the customer in one sentence at the end."

If your team is adjusting to AI tools generally and not just for prompts like these, it is worth looking at how managers are handling the shift in AI training for managers leading a team that works with AI, because the friction is rarely the tool itself, it's getting a team to trust the output enough to stop double-checking everything.

Work with me

Want AI doing the heavy lifting in your marketing?

I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.

Where Claude falls short for PMs

Claude is weaker than some alternatives at anything requiring live data, like current market sizing or competitor pricing, because its knowledge has a cutoff and it won't reliably browse the way some other tools do. For that kind of research, you're better off treating Claude as the writer and editor, and doing the fact-finding somewhere else first, then feeding the facts in. If your product work overlaps with coding or technical specs that need to go straight to engineers, it's also worth comparing how Claude's coding-specific tooling stacks up against alternatives in Claude Code vs Cursor, and if your team uses OpenAI's agent for implementation work, Codex prompts for coding briefs covers the handoff from spec to code in more detail.

A prompt for user research synthesis

Say you've just finished eight customer interviews about a pricing change and you have raw, unedited transcripts, roughly 25,000 words combined. This is where Claude's long context matters:

  • "Here are transcripts from eight customer interviews about our proposed pricing change [paste all eight]. Pull out every direct quote that expresses concern or hesitation, grouped by theme. Then pull out every quote that expresses enthusiasm. Don't summarise the quotes, use the exact wording. Tell me which theme came up most often across different customers, not just most frequently overall."

That distinction at the end, between a theme repeated by one chatty customer versus one raised by several different people, is exactly the kind of nuance that separates a useful synthesis from a misleading one, and it's the sort of instruction most people forget to include.

A quick note on hiring and skills

If you're a PM hiring for your team, or a PM job hunting yourself, the same prompting discipline applies to writing applications. Specificity over fluff wins in both directions. If that's relevant to you right now, Claude prompts for job applications and cover letters and Gemini prompts for job interviews cover that side of things.

Frequently asked questions

Is Claude better than ChatGPT for product managers?

Claude tends to handle long documents like transcripts and sprawling backlogs better without losing detail, while ChatGPT has broader plugin and browsing support. Most PMs benefit from using both rather than picking one permanently.

Can Claude write an entire PRD on its own?

No, not a good one. Claude can structure and tighten a PRD brilliantly if you give it your raw notes, decisions and constraints, but it has no access to your data, customers or company context, so decisions still need to come from you.

What's the biggest mistake PMs make with Claude prompts?

Asking Claude to make prioritisation or roadmap decisions instead of asking it to organise information so you can decide. The tool is excellent at structure and terrible at judgement calls it has no data for.

How long should a Claude prompt for product work be?

Longer than you'd think for anything involving judgement, short for anything routine. A stakeholder update prompt can be two sentences. A PRD or research synthesis prompt should specify sections, length, tone and what to flag, usually four or five sentences.

Run a business, a brand or a blog? I feature tools, products and expert guest posts for an audience of 15,000. See how it works

Published and maintained by the Lilach Bullock team, covering marketing, AI and business growth.
Your buyers are asking AI who to use. Does it say you?

See for free whether ChatGPT, Claude, Perplexity, Gemini and Google name you, and get the plan to become the answer.

Check my AI visibility →
Sundays only

Get the Sunday newsletter.

One email a week. AI experiments, marketing tactics, and the workflows Lilach is building right now in her own business.

Subscribe free

Let’s get your marketing running on AI.

Book a free 30-minute call

We figure out what you need, where AI fits in, and what working together would look like.

Book the call →

Or take the 30-second calculator

You’ll see the hours and the money quietly leaking out of your week, and the three workflows worth building first.

Take the calculator →

Or grab the free AI resource library

Prompt packs, templates, checklists, and swipe files. The exact tools I build for paying clients. Yours, free.

Get the library →
Keep reading

More from the blog.