# **The Founder Board: Test Your Business in Claude Before You Spend a Dollar** *By the end of this guide you will have 11 AI advisors running your business on paper: a board, a CFO, a marketing lead, and a panel of 100 simulated buyers that tells you who says yes, who walks away, and why.* **Want more AI workflows like this?** Join our free community for the best AI tools, prompts, and shortcuts. [**Join here: skool.com/viral-ads**](https://www.skool.com/viral-ads) ## --- **What this actually is** **The short version:** You give Claude eleven jobs, one for each seat in a company. Each seat looks at your idea from one angle, writes its verdict down, and the next seat builds on it. Think of it as a dress rehearsal for your business. Actors rehearse so the mistakes happen in an empty theatre, not on opening night. This is the same thing for your money. * **Board of Directors:** three lenses score your idea and vote to fund it or pass. * **Competitor Scout:** who you are up against, what they charge, and what their customers hate. * **Consumer Panel:** 100 simulated buyers, each with a different income, habit and doubt. * **Pricing Strategist:** the price range buyers will actually accept. * **Offer Architect:** bonuses, guarantee and urgency that answer the objections. * **CFO:** your real profit per sale, break-even per day, and the cash you need. * **Marketing Director:** where your buyers are and a dated 30-day campaign. * **Brand Designer:** names, voice and look, with the checks to run. * **Operations Manager:** suppliers, permits, routines and risks. * **Launch Lead:** a cheap real-world test first, then the launch plan. * **Business Planner:** one plan with a verdict computed from the numbers. **WHY THIS EXISTS** | Most businesses find out the price is wrong, the margin is thin or the buyer does not care only after the lease is signed and the ads are paid for. This flips the order. You find the holes on paper, while they still cost nothing. | | :---- | ## --- **Where this pack came from (and why there is no download link)** The pack in the video is a brand-new community project on GitHub, published in October 2026 under a free license. The idea is great. It is also only a few weeks old, has about 1,500 stars, and runs small Python scripts on your computer. **Our rule:** we only point you to GitHub code with 5,000+ stars and a public track record, or to an official company source. A GitHub star is like a bookmark from a developer, so a high count means thousands of people have looked at the code. "Probably fine" is not good enough when your files are on the line. **THE GOOD NEWS** | The value was never the code. It is the method: eleven seats, a fixed order, and a buyer panel. You get a clean version you own: nothing to download, nothing to run, and it works in the normal Claude app, even on the free plan. | | :---- | ## --- **The thing nobody tells you** **AI BUYERS ARE TOO NICE** | AI chatbots like Claude like to please you, so a simulated panel says yes more often than real people do. Read the yes rate as a ceiling, never a forecast. The gold is in the no answers: the exact objections, in the buyer's own words, and what would flip them. | | :---- | **The order matters too.** Each seat feeds the next one. Pricing needs the panel's price answers. The offer needs the panel's objections. The CFO needs the price. Run them out of order and you get generic advice instead of advice about your business. | Board \> Competitors \> Consumer Panel \> Pricing \> Offer \> CFO \> Marketing \> Brand \> Ops \> Launch \> Plan | | :---- | ## --- **Set it up in 5 minutes (no code)** ### **Option A: a Claude Project (best for most people)** A Project is a folder inside Claude. Every chat in it follows the same instructions and can read the same files, so your advisors remember your business between conversations. 1. **Open Claude** Go to [claude.ai](https://claude.ai) or open the Claude desktop app and sign in. Projects work on the free plan (up to 5 projects). 2. **Create the project** Click **Projects** in the left sidebar, then **\+ New Project**. Name it "Founder HQ". 3. **Paste the master instructions** Click **Set project instructions**, paste the full FOUNDER HQ block from the master instructions section below, and save. This is what turns Claude into your eleven advisors. 4. **Add your real facts (optional, but it makes a big difference)** Use the **\+** button in the project knowledge area to add anything real: a rough budget, supplier quotes, competitor links, a price list. 5. **Start your first chat inside the project** Type the Seat 1 prompt from this guide. That is it. You are running a board meeting. **PRO MOVE** | After each seat finishes, copy its answer, then in the project knowledge area click \+, choose Add text content, paste it, and name it after the seat (for example seat-3-panel). Every later chat can then read it. | | :---- | ### **Option B: real skills in Claude Code (for people who already use it)** A skill is a saved instruction file that Claude loads on its own when the job matches, like an app on your phone that opens when you need it. In Claude Code (Anthropic's coding tool that runs in your terminal), skills live in a folder on your computer and become slash commands. Paste this into Claude Code: | Turn the FOUNDER HQ instructions below into 11 Claude Code skills. For each seat, create the folder \~/.claude/skills/founder-\/ with one SKILL.md file. Each SKILL.md starts with YAML frontmatter: a name (lowercase, hyphens only) and a description (one sentence saying when to use this skill). In the body, copy the GROUND RULES plus that seat's method. Every skill saves its result to a founder/ folder in my current project and reads the earlier seats' files from that folder first. Do not download anything and do not write or run any scripts. When you are done, list the 11 slash commands and the order to run them. \[paste the whole FOUNDER HQ block here\] | | :---- | **What you get:** eleven commands such as /founder-board and /founder-consumer, and every result saved as a file your other seats read automatically. ### **Option C: skills in the Claude app** 1. **Turn on code execution** Settings, then **Capabilities**, then switch on code execution and file creation. Skills need it. 2. **Ask Claude to package all 11 seats** In a new chat, paste the prompt below with the FOUNDER HQ block under it. 3. **Download the zips** Claude hands you 11 zip files. Click each one to download it. 4. **Upload them one by one** Go to **Customize**, then **Skills**, click **\+**, choose **Create skill**, then **Upload a skill** and pick a zip. Repeat for all 11 and toggle each one on. | Package the FOUNDER HQ instructions below as 11 separate Claude skills, one per seat. Names: founder-board, founder-competitors, founder-consumer, founder-pricing, founder-offer, founder-cfo, founder-marketing, founder-brand, founder-ops, founder-launch, founder-plan. Each skill is a folder with that exact name containing one SKILL.md file. The file starts with a name (same as the folder) and a one-sentence description of when to use it, then the GROUND RULES, then that seat's method. No scripts, no other files. Zip each folder separately and give me the 11 zip files to download. \[paste the whole FOUNDER HQ block here\] | | :---- | **WHICH OPTION SHOULD YOU PICK?** | Option A gives you the same eleven advisors with zero setup. Pick Option C only if you want Claude to load a seat on its own whenever the topic comes up, in any chat. | | :---- | ## --- **The master instructions (paste this into your project)** This is the whole engine. Every seat's method is written out, so Claude does the full job every time instead of improvising. | You are FOUNDER HQ: eleven advisors who stress-test my business idea before I spend money. Each advisor is a "seat". When I say "Run seat N" (or the seat name), act ONLY as that seat. GROUND RULES (every seat) \- Ask me for missing facts before you guess. If you must estimate, label it ESTIMATE. \- When you research, use real public sources and give the link for every fact.   Never invent competitors, prices, reviews or quotes. \- Read what earlier seats wrote (in this chat or in the project files) and build on it. \- End every seat with: the 3 biggest risks, the 3 next actions, and a one-line verdict. \- Plain English. Short sentences. Use tables when you compare things. \- You are not a lawyer, accountant or financial advisor. Flag anything that needs one. SEAT 1: BOARD OF DIRECTORS Three board members. Each applies one published framework. Never speak as the real person.   A. Offer lens (from Alex Hormozi's book \$100M Offers): dream outcome, how likely it feels,      how fast it arrives, how much effort it costs. Is it so good people feel stupid saying no?   B. Monopoly lens (from Peter Thiel's book Zero to One): what do we know that others don't?      Can we own a small market first? What makes us 10x better, not 10% better?   C. Product lens (from Walter Isaacson's biography Steve Jobs): focus, simplicity, the whole      experience end to end. What should we cut? Each member scores the idea 1 to 10, names what would kill it, and votes FUND, FUND IF, or PASS. Turn every FUND IF condition into a checklist the other seats must answer. SEAT 2: COMPETITOR SCOUT Map direct competitors, indirect ones, and substitutes (what people do instead of buying). For each: price, positioning, and what their customers complain about in public reviews. Link every fact. Finish with the gap we can own. SEAT 3: CONSUMER PANEL (100 simulated buyers) 1\. Build the target customer with me: who, age, income range, where they buy today, what they spend. 2\. Deal 100 buyer cards. Mix income (low, middle, high), habits (heavy, occasional, never buys    this category) and one main objection each (price, trust, time, already has a solution,    doubts new brands, and so on). Show the mix as a table so I can check it is realistic. 3\. Put each buyer in one specific buying moment (scrolling Instagram, comparing on Google,    standing in the store, a friend recommended it) and show them my pitch. 4\. Each buyer answers: BUY, MAYBE or NO; the reason in their own words; what would flip a NO;    and four price answers: too cheap to trust, a bargain, getting expensive, too expensive. 5\. Tally the yes, maybe and no overall and by segment, rank the top 5 objections, list the    5 best phrases buyers used, and name the segment most likely to buy first. Run the panel in batches of 25 and keep a running tally. Remind me that simulated buyers lean agreeable, so the yes rate is a ceiling, not a forecast. SEAT 4: PRICING STRATEGIST Use the panel's four price answers (the Van Westendorp method) to find the acceptable price range and the sweet spot. Compare it with competitor prices and my costs. Recommend one price to test and one backup price. SEAT 5: OFFER ARCHITECT Answer the top objections with the offer: bonuses, a guarantee, honest urgency (a real deadline or a real limit, never fake), and a name for the offer. Then re-run a quick 30-buyer panel on the new offer and show whether the yes rate moved and why. SEAT 6: CFO Unit economics in a table: price, cost per unit, payment fees, contribution per sale (price minus every per-sale cost), fixed costs per month, break-even sales per month and per day, profit margin at my planned volume, year 1 month by month with a slow ramp, the cash I need before it pays for itself, and the payback month. Then three what-ifs: price down 10%, costs up 15%, sales down 30%. Show every formula. Mark every estimate. SEAT 7: MARKETING DIRECTOR Positioning in one sentence. The 3 channels where these buyers already spend time. The most I can pay to win one customer (from the CFO numbers). A dated 30-day launch campaign. Ten hooks written in the buyers' own words from the panel. SEAT 8: BRAND DESIGNER Ten name ideas, each with the checks to run (trademark search, .com domain, Instagram handle). A voice in three words, a one-line promise, and a short brief for the look (colors, type, feel). SEAT 9: OPERATIONS MANAGER Suppliers and tools, staffing, day-one routines, the permits and licenses to check for my location, and a risk register (risk, likelihood, impact, fix). Send real costs back to the CFO. SEAT 10: LAUNCH LEAD Start with the cheapest real-world test (a pre-sale page, a pop-up, 10 sales calls, a small ad test) and say what result counts as a pass. Then a dated countdown, a launch-day run sheet, and the plan for the first 30 days. SEAT 11: BUSINESS PLANNER Compile every seat into one plan. The verdict is computed, not felt:   PROFITABLE \= every sale makes money AND year 1 ends in profit AND break-even fits my capacity                AND at least 30% of the panel said BUY.   NOT YET \= any one of those fails. Say which one and what would fix it.   INCOMPLETE \= a seat is missing. Say which one. | | :---- | **Why it is built this way:** the ground rules stop the three classic AI failures: guessing your numbers, inventing competitors, and forgetting what the last seat said. The computed verdict in seat 11 stops the last failure, which is you falling in love with your own idea. ## --- **The 11 seat prompts (run them in this order)** Fill in the brackets with your real details. The more specific you are, the sharper every seat gets. **WEAK VS STRONG** | Weak: busy moms. Strong: moms in Austin, 30 to 42, household income \$90K to \$150K, who order takeout 3 times a week. | | :---- | ### **Seat 1: Board of Directors** | Run seat 1: Board. My idea: \[what you sell\], for \[who\], at \[price\], and they will find me through \[channel\]. My budget to start is \[amount\]. | | :---- | **Why it works:** Three different lenses catch three different ways an idea dies: a weak offer, a crowded market, or a product that tries to do too much. **You get back:** Three scores out of 10, three kill risks, three votes, and a FUND IF checklist that the rest of the seats have to answer. ### **Seat 2: Competitor Scout** | Run seat 2: Competitor Scout. I sell in \[city or online\]. Here are the competitors I already know: \[names or links\]. | | :---- | **Why it works:** The fastest way to find your angle is to read what people hate about the options they already pay for. **You get back:** A table of competitors with prices and real complaints, every row linked, plus the one gap you can own. ### **Seat 3: Consumer Panel** | Run seat 3: Consumer Panel. My target customer: \[age\], \[income\], \[where they live\], \[what they buy today and how often\]. My pitch in two lines: \[pitch\]. | | :---- | **Why it works:** One hundred different people reacting to the same pitch shows you patterns no single opinion can: which segment says yes, and the exact words behind every no. When it stops at buyer 25, type: continue from buyer 26\. **You get back:** A 100-row panel, yes, maybe and no counts by segment, the top 5 objections ranked, and the phrases buyers used, which become your ad copy later. ### **Seat 4: Pricing Strategist** | Run seat 4: Pricing. Use the panel price answers. My cost per unit is \[amount\]. | | :---- | **Why it works:** Most founders pick a price by gut. Four simple questions about when a price feels too cheap or too expensive find the range where buyers stop thinking it is cheap and start thinking it is too much. **You get back:** An acceptable price range, a sweet spot, a comparison with competitors, and one price to test first. ### **Seat 5: Offer Architect** | Run seat 5: Offer. Kill the top 3 objections from the panel, then re-test on 30 buyers. | | :---- | **Why it works:** You rarely need a new product. You need an offer that answers the reasons people walk away. **You get back:** An offer stack (bonuses, guarantee, honest deadline, a name) and a before and after yes rate. ### **Seat 6: CFO** | Run seat 6: CFO. My fixed costs per month: \[rent, staff, software, other\]. My cost per unit: \[amount\]. Planned sales per day: \[number\]. Cash I have: \[amount\]. | | :---- | **Why it works:** Revenue feels good, but contribution (what each sale leaves after its own costs) pays the bills. This seat shows how many sales a day you need just to stay alive. **You get back:** A full unit economics table, break-even per day, year 1 month by month, the cash you need, and three what-if scenarios. ### **Seat 7: Marketing Director** | Run seat 7: Marketing. My launch date is \[date\]. My monthly marketing budget is \[amount\]. | | :---- | **Why it works:** The CFO tells you the most you can pay for one customer, so the plan never spends more than a customer is worth. **You get back:** A one-line position, 3 channels, a dated 30-day campaign, and 10 hooks in your buyers' own words. ### **Seat 8: Brand Designer** | Run seat 8: Brand. Words I like: \[3 words\]. Brands I admire: \[names\]. | | :---- | **Why it works:** A name that is taken, or a handle you cannot get, costs real money to fix after launch. **You get back:** Ten names with the checks to run, a voice, a one-line promise, and a brief for the look. ### **Seat 9: Operations Manager** | Run seat 9: Ops. My location is \[city, state or country\]. I will run this \[alone or with N people\]. | | :---- | **Why it works:** Permits, suppliers and staffing are where cash quietly leaks. This seat sends those costs back to the CFO. **You get back:** A supplier and tool list, day-one routines, permits to check, and a risk register. ### **Seat 10: Launch Lead** | Run seat 10: Launch. Design the cheapest real-world test I can run in 7 days. | | :---- | **Why it works:** This is where the simulation meets real people. A pre-sale page or ten sales calls tell you more than any AI panel. **You get back:** A test with a clear pass mark, a dated countdown, a launch-day run sheet, and a 30-day plan. ### **Seat 11: Business Planner** | Run seat 11: Plan. Compile every seat and give me the verdict. | | :---- | **Why it works:** A verdict based on rules stops you from falling in love with your own idea. **You get back:** One business plan with a verdict of Profitable, Not yet, or Incomplete, and the exact reason. ## --- **What the CFO seat actually calculates (worked example)** Here is the math in plain English, so you can sanity-check what Claude gives you. Say you want to open a matcha bar and sell iced lattes at \$6.50. * **Cost per cup:** matcha, milk, ice and cup come to about \$1.40. Card fees are about 3%, so roughly \$0.20. * **Contribution per cup:** \$6.50 minus \$1.40 minus \$0.20 \= **\$4.90**. This is what each sale leaves to pay the fixed bills. * **Fixed costs per month:** rent \$6,500, staff \$6,000, everything else \$1,500 \= **\$14,000**. * **Break-even:** \$14,000 divided by \$4.90 \= about 2,858 cups a month, or **about 96 cups every day** just to stay at zero. **ONE NUMBER DECIDES IT** | If your street realistically gives you 60 customers a day, this business is dead on paper, and you just saved yourself a lease. If it gives you 250, you have room. One number decides it, and most people never calculate it before they sign. | | :---- | These example numbers are made up to show the method. Your CFO seat will ask for your real ones. ## --- **Get it right and stay safe** **THE PANEL IS NOT YOUR CUSTOMERS** | Simulated buyers are research, not proof. Never quote them as testimonials or reviews in your marketing. The FTC (the US consumer protection agency) rule on fake reviews and testimonials (2024) treats made-up customer quotes as deceptive, and the fines are per violation. | | :---- | **THE BOARD IS NOT THOSE PEOPLE** | The board applies ideas from published books by Alex Hormozi, Peter Thiel and Walter Isaacson. It does not speak for them, and none of them endorse your plan. Do not say "Hormozi approved my offer" anywhere. | | :---- | **NUMBERS ARE ONLY AS GOOD AS YOUR INPUTS** | If Claude marks a number as an estimate, get a real quote before you rely on it. Check legal structure, payroll and tax with an accountant. This guide is not financial, legal or tax advice. | | :---- | **Keep private data out.** Do not paste real customer lists, card details or anyone's personal information into the panel. Describe your customer as a type, not as named people. ## --- **Pro moves most people miss** * **Feed the scout real complaints.** Paste 20 one-to-three-star reviews of your top competitor into seat 2\. Their pain is your positioning. * **Test two pitches, compare the gap.** Run the panel on version A and version B. The absolute yes rate is inflated, but the difference between the two is useful. * **Ask for the near-misses.** "Which 10 buyers were closest to yes, and what single change flips them?" That one question often rewrites the offer. * **Make the board argue.** "Have the Offer lens and the Product lens debate their biggest disagreement, then tell me who is right and why." * **Steal the buyers' words.** The panel phrases are written the way customers talk. Use them as ad hooks and headline tests. * **Close the loop with real people.** Take the top 3 objections into 10 real conversations. If real people raise the same ones, the panel was right. ## --- **Common mistakes (and how to avoid them)** * **Treating the yes rate as a sales forecast.** Treat it as an upper limit. Real buyers say yes less. * **Letting Claude guess your costs.** Give it real quotes. A wrong rent number breaks every later seat. * **A vague customer.** "Everyone" gives you a panel of nobody. Pick one clear type of buyer first. * **Asking for 100 buyers in one reply.** Long answers can stop halfway. Run batches of 25 and say "continue from buyer 26". * **Skipping seat 10\.** The simulation is the rehearsal. The cheap real-world test is the show. ## --- **Your first week** 1. **Day 1** Create the Founder HQ project, paste the master instructions, run seat 1\. 2. **Day 2** Run seat 2 (competitors) and seat 3 (the 100-buyer panel). 3. **Day 3** Run seat 4 (pricing) and seat 5 (offer), including the 30-buyer re-test. 4. **Day 4** Collect real quotes for your costs, then run seat 6 (CFO). 5. **Day 5** Run seat 7 (marketing) and seat 8 (brand). Check that the names you like are actually free. 6. **Day 6** Run seat 9 (operations) and seat 10 (launch). Pick your cheap real-world test. 7. **Day 7** Run seat 11 for the verdict. If it says Profitable, start the real test this week. ## --- **Resources and Links** * [Claude (web app)](https://claude.ai) * [How to create and manage Claude Projects (official help)](https://support.claude.com/en/articles/9519177-how-can-i-create-and-manage-projects) * [Using skills in Claude (official help)](https://support.claude.com/en/articles/12512180-using-skills-in-claude) * [Skills in Claude Code (official docs)](https://code.claude.com/docs/en/skills) * [USPTO trademark search (for seat 8\)](https://tmsearch.uspto.gov) * [SBA guide to business licenses and permits (for seat 9\)](https://www.sba.gov/business-guide/launch-your-business/apply-licenses-permits) * [Alex Hormozi books, including \$100M Offers (Acquisition.com)](https://www.acquisition.com/books) * [FTC Consumer Reviews and Testimonials Rule: questions and answers](https://www.ftc.gov/business-guidance/resources/consumer-reviews-testimonials-rule-questions-answers) --- **Want more AI workflows like this?** Join our free community for the best AI tools, prompts, and shortcuts. [**Join here: skool.com/viral-ads**](https://www.skool.com/viral-ads)