The best Claude prompts for founders

By the Chat Judge team · Updated 2026-10-19 · 5 min read · Links go to prompts in our library; creators are credited on each prompt.

In short: Claude is at its best on long, structured work where you want it to push back: pricing, specs and reviews of something you already wrote. This guide covers the prompts in the library that were written with that in mind, from designing tiers to handing a scoped change to Claude Code.

Pricing, where you want the model to argue with you

The most useful thing a model can do with pricing is refuse to bless a number you made up. Design pricing tiers asks for tiers with names, limits, prices and upgrade fences, and it tells Claude to gate on capacity rather than on the feature that drives the customer's revenue. It also asks Claude to challenge you if your numbers look anchored on your own willingness to pay. That only works if you paste real cost drivers and any benchmarks you actually have.

Unit economics check is the follow-up. You give per-request costs and volumes and ask which usage types are structurally unprofitable, and which cost levers would need a quality check before you pull them. Counter-offer to a discount request and Pilot pricing for an existing customer handle the two conversations that follow: what to give, hold and ask for when someone wants a discount, and how to price a pilot without breaking current revenue. None of these produce a price you can trust on their own. They produce a structure and a list of questions your numbers have not answered.

Handing work to Claude Code

Claude Code is a coding agent, and it does better with one precise document than with a chat transcript. Consolidated PRD for a Claude Code handoff asks Claude to combine everything already decided into a single markdown file: problem, locked decisions, architecture, data model, milestones ordered by risk, and acceptance criteria, with open questions separated out so they are not mistaken for decisions.

For a smaller change, Paste-ready Claude Code prompt writes one prompt you can paste directly, referencing PRD sections instead of restating them and saying explicitly what not to build. Before you build at all, Review a PRD or bug report looks for what a spec misses or conflates and ranks the fixes by cost and quality impact. Setup or integration instructions for Claude Code splits the steps only you can do in a browser from the ones the agent can run, and tells it to check the current README before trusting an install command.

Decks and pages, reviewed rather than written

When Claude is writing a deck, the constraint that matters is honesty. Sales deck from scratch takes a fixed structure, problem through to call to action, and restricts it to claims that are true today, flagging any slide where you have not supplied proof. Review an existing deck, bluntly is for the deck you already have: what is unclear, what contradicts another slide, and what a procurement or investor reader will attack. Competitor battle card builds a comparison against named competitors and asks for an honest note on where they lead, with every competitor statistic marked by source and date so you can verify it before a deal.

The same review habit helps on the site. Critique a live page goes section by section and ranks the top three changes, and Pricing page copy writes tier cards and an FAQ while calling out anything in the tier structure that will surprise people at renewal.

Outreach and negotiation

Cold outreach to a senior target keeps a first message under 120 words, led by the recipient's problem rather than your feature list, and asks Claude to tell you if it still reads like a vendor pitch. Negotiation reply separates the thinking from the writing: first what to give, hold and ask for, and any inconsistencies in your own proposal, then a reply under 200 words. Read both before you send either. Claude does not know your walk-away point unless you wrote it down.

Where founders waste Claude

The expensive failure mode is pasting a whole strategy and asking for "thoughts". You get a flattering summary. Constrain the job instead. Feasibility and go/no-go maps what you have against what a tender requires and names the rules that could disqualify you. Compare two documents clause by clause sorts every difference between a template and a filled-in version into material, naming and cosmetic, which is exactly the check you want before signing. The founder context block at the top of the chat tells Claude who you are and which decisions are locked, so it argues with the work instead of relitigating your strategy.

For the site, pair the review prompts with Choose a domain name and plan the migration when you are renaming: it checks collisions and lists the unglamorous steps, from DNS to redirects to canonical tags. And Forwardable warm-intro blurb is the short Claude prompt worth keeping, because introductions stall when your contact has to rewrite them.

How to get good results from any of these prompts

Fill in every placeholder. Anything in [BRACKETS] is a fact only you know. Leave one empty and the model will guess.
Tell it what not to invent. Add "use only the facts above, and flag anything missing" if the prompt does not already say it.
Ask for one version first. Fix the direction, then ask for variations.
Check every name, number and promise before anything leaves your hands.
Edit it into your own voice. Read it aloud once; if it does not sound like you, change it.

Prompts in this guide

PricingClaudeDesign pricing tiersSetting or resetting pricing for a product with real cost drivers.View prompt PricingClaudeUnit economics checkTesting whether your pricing survives your real usage costs.View prompt PricingClaudeCounter-offer to a discount requestA customer wants a discount and you need a position before you reply.View prompt PricingClaudePilot pricing for an existing customerMoving a current customer onto a new product without disrupting revenue.View prompt ProductClaudeConsolidated PRD for a Claude Code handoffAfter a long design conversation, to turn every decision into one build-ready document.View prompt ProductClaudePaste-ready Claude Code promptHanding one scoped change to a coding agent, not a whole document.View prompt ProductClaudeReview a PRD or bug reportBefore building, to find what a spec misses or conflates.View prompt ProductClaudeSetup or integration instructions for Claude CodeGetting a coding agent to connect or deploy something, with only the browser steps left for you.View prompt SlidesClaudeSales deck from scratch (only claims you can prove)Building a first sales deck for a meeting where you want a pilot or next step.View prompt SlidesClaudeReview an existing deck, bluntlyBefore a deck goes to procurement, a customer or investors.View prompt SlidesClaudeCompetitor battle cardArming sales for a deal against named competitors, without fabricating stats.View prompt EmailsClaudeCold outreach to a senior targetReaching a senior person you do not know, with one strong proof point.View prompt EmailsClaudeNegotiation reply: what to give, hold and askA customer or partner has pushed back on price or terms and you must reply.View prompt WebsitesClaudeCritique a live page like a conversion copywriterYou have a live page that is not converting and want a ranked fix list.View prompt WebsitesClaudePricing page copy that avoids renewal surprisesWriting or fixing a pricing page and wanting to catch churn traps in the tier structure.View prompt

Questions people ask

Do I need Claude Code for the spec prompts?

No. They are written for Claude Code, but the same structure helps any coding assistant or a human engineer.

Can Claude tell me what to charge?

No. It does not know your costs or customers. The pricing prompts structure the decision and push back on weak assumptions; the number is still your call.

Will these work in ChatGPT too?

Yes. They are plain text. They were written with Claude in mind because they ask for long, structured answers and critique.

Keep exploring

All Claude prompts Pricing & deal strategy Product specs & Claude Code handoffs Slides & presentations Emails Website & product copy

More guides

Written by the Chat Judge team from prompts in our library. We do not run model benchmarks, and nothing here is a claim about results you will get. Spot a problem? Tell us or submit a better prompt.