9 Claude Prompts That Generate Startup Ideas Like Shark Tank Investors
Summary
-
Claude Opus 4.6 can simulate a full Shark Tank investor panel to generate and rank fundable startup ideas based on your actual budget, skills, and time constraints
-
The nine-prompt sequence covers every stage of the startup process from idea generation and market validation to pitch decks, cold outreach, and a 30-day post-launch sprint
-
Each prompt builds on the previous one, so by the end you have a complete founder’s document covering financials, positioning, launch hooks, and a reality-check audit
-
Running the full sequence takes roughly three hours and works best when treated as separate sessions rather than one sitting
-
The prompts are designed to kill bad ideas fast and sharpen good ones before you spend any money building
I have sat on more business ideas than I care to admit. The cycle is always the same: get excited, start overthinking, wonder if the market is too crowded, and eventually talk myself out of it before I even start.
If that sounds familiar, you are not alone.
What changed things for me was running my ideas through Claude before spending any money or time building.
These nine prompts are designed to put your concept in front of a simulated panel of investors, the kind of brutal, numbers-driven scrutiny that most founders only experience after they have already made costly mistakes.
I use them to stress-test ideas fast, and they have saved me from chasing more than a few dead ends.
The prompts work best with Claude Opus 4.6, which handles the kind of complex, multi-step reasoning these tasks demand.
I go through all nine in sequence because each one feeds into the next. By the end, you will have a startup plan that covers everything from market validation to your first 50 customers.
You do not need a business background to follow along. You need an idea, about three hours, and the willingness to hear honest feedback.
Let me show you exactly how I use each prompt.
What Is the Shark Tank Prompt Framework and Why Does It Work
Most AI prompts for business ideas are too vague. You ask for startup ideas and you get a generic list that could have come from a 2019 blog post.
What makes this framework different is that it forces Claude to think like an investor, not a brainstormer.
Each prompt assigns Claude a specific expert role, gives it clear constraints to work with, and demands structured, numbered output.
That combination pushes the model to go deeper than surface-level suggestions. Instead of “start a subscription box business,” you get a named concept, a quantified problem, a revenue model, and a brutally honest verdict on whether it would actually get funded.
I have run dozens of ideas through this framework, and the pattern is consistent: bad ideas get killed in the first two prompts, and the ones worth pursuing become sharper and more specific with every step.
That is the point. You want to fail fast on paper, not after six months of building.
1. The Shark Tank Idea Generator
This is where everything starts. Prompt 1 turns Claude into a combined voice of all five Shark Tank investors and asks it to generate eight startup ideas based on your personal constraints.
It is the most important prompt in the sequence because it sets the direction for everything that follows.
Here is the full prompt to copy and use:
You are now a panel of Shark Tank investors (Mark Cuban + Lori Greiner + Kevin O’Leary + Daymond John + Barbara Corcoran) combined into one brutally honest AI. Generate 8 startup ideas that would actually get funded in 2026. My constraints: Budget to start: [your budget]. Time I can commit per week: [hours]. Skills I bring: [your skills]. Niche I am passionate about: [optional]. For each idea, provide: a one-sentence company name and tagline, the problem it solves with a quantified pain point, a simple solution in one to two sentences, the target customer profile, the revenue model, why now in 2026, the competitive edge, realistic first-year revenue potential, a one-sentence investor hook, and a brutal verdict of invest or pass with a reason. Rank them from most fundable to least.
What I do before running this prompt is spend five minutes filling in my actual constraints honestly.
The more specific I am, the more useful the output becomes.
Here is an example of how I fill it in:
| Field | Example Input |
|---|---|
| Budget to start | $500 to $5,000 |
| Hours per week | 15 hours |
| Skills I bring | Content writing, SEO, social media |
| Niche I care about | AI tools and productivity |
When Claude returns the eight ideas, I do not pick my favourite. I pick the one with the strongest investor verdict and the most specific revenue logic. That idea goes into Prompt 2.
2. The Market Validation Sniper
Once I have a winning idea from Prompt 1, I do not move to building anything. I run it through this prompt first.
Prompt 2 assigns Claude the role of a VC analyst who has reviewed over a thousand pitches, and it asks for a full market validation report on the idea I have chosen.
This is the prompt that saves the most money.
I have had ideas that looked great on paper get completely dismantled here because the market was either too small, too saturated, or moving in the wrong direction. Better to find that out now.
Here is the full prompt:
You are a VC analyst who has evaluated 1,000+ pitches. Take this idea: [paste your chosen idea from Prompt 1]. Run a real-time market validation covering: TAM, SAM, and SOM estimates using the latest available data, the top 3 to 5 competitors with their strengths and fatal flaws, where the market is currently underserved, search volume and trend signals, customer pain language from Reddit, X, and TikTok, a quick validation test I can run in 24 to 48 hours, a one-sentence proof that this market is growing, and a final verdict of green light, pivot needed, or kill it. Be numbers-driven and brutally honest.
The output from this prompt gives me a structured picture I can actually act on.
Here is what a typical validation breakdown looks like across the key sections:
| Validation Area | What Claude Delivers |
|---|---|
| Market size (TAM/SAM/SOM) | Dollar estimates with reasoning |
| Competitor analysis | 3 to 5 named competitors, strengths, and fatal flaws |
| Gap analysis | Where the market is underserved right now |
| Trend signals | Search volume direction and keyword difficulty |
| Customer sentiment | Pain language pulled from social platforms |
| Quick test | A 24 to 48-hour poll, landing page, or DM script |
| Final verdict | Green light, pivot needed, or kill it |
If Claude returns a “pivot needed” verdict, I do not scrap the idea entirely. I take the gap analysis section and ask myself whether I can reposition the concept to fill that gap.
Sometimes one pivot unlocks a much stronger angle than the original idea.
3. The One-Page Business Model Canvas
By this point, I have an idea that has survived investor scrutiny and market validation.
Prompt 3 is where I turn that idea into a structured business model. Claude takes on the role of a lean startup expert and fills out a complete Business Model Canvas in text form.
I use this prompt because it forces clarity on the parts of a business plan that founders most often skip: how customers will actually find the product, what the cost structure looks like month to month, and what the unfair advantage is.
Those three things alone determine whether a business survives its first year.
Here is the prompt:
You are a lean startup expert who turns ideas into one-pagers investors love. Using this idea and market data: [paste key outputs from Prompts 1 and 2]. Fill out a complete Business Model Canvas covering: customer segments, value proposition with a one-sentence promise and 4 to 6 bullet benefits, channels for discovery and purchase, customer relationships, revenue streams with price points, key resources needed, key weekly and monthly activities, key partnerships that can accelerate growth, cost structure with fixed and variable monthly burn, and a one-sentence unfair advantage. End with a one-paragraph summary of why this business wins.
The canvas output is one of the most useful documents in the entire workflow because it doubles as a reference I return to throughout the build.
I keep it open alongside Prompt 4 when I am working on projections so the numbers stay grounded in the actual business model.
Here is what the nine canvas blocks look like when Claude fills them out for a content-based AI tools business as an example:
| Canvas Block | Example Output |
|---|---|
| Customer Segments | Solopreneurs and small teams using AI tools daily |
| Value Proposition | Save 5 hours a week with curated AI workflows |
| Channels | SEO, YouTube Shorts, newsletter |
| Customer Relationships | Self-serve with community support |
| Revenue Streams | $19/month subscription, $97 one-time course |
| Key Resources | Content library, AI tool access, email list |
| Key Activities | Weekly content, prompt testing, list growth |
| Key Partnerships | AI tool companies, affiliate networks |
| Cost Structure | $200/month fixed, $50 variable per 100 subscribers |
| Unfair Advantage | First-mover content depth on emerging AI tools |
4. The 12-Month Revenue and Growth Projection Model
This is the prompt most founders avoid because it forces you to put real numbers on paper.
Prompt 4 turns Claude into a financial modeler and asks it to build a month-by-month revenue projection based on assumptions you provide.
The assumptions are the key part. The more honest I am here, the more useful the output becomes.
I always run this prompt twice: once with conservative assumptions and once with aggressive ones.
The gap between those two scenarios tells me exactly how much my results depend on execution versus market conditions. That is information worth having before I spend a single dollar.
Here is the full prompt:
You are a financial modeler who builds investor decks. Using this business: [paste canvas summary from Prompt 3]. Create a realistic 12-month projection using these assumptions: Launch month: [month]. Monthly customer acquisition goal: [number]. Conversion rate: [percentage]. Average order value or subscription price: [dollar amount]. Monthly churn: [percentage]. CAC target: [dollar amount]. Deliver a month-by-month table covering customers, revenue, costs, and profit. Include cumulative revenue at month 12 across conservative, base, and aggressive scenarios, the breakeven month, funding needed and runway, a one-sentence summary of what drives the target revenue by what month, and the biggest risk to the numbers with a mitigation strategy.
The month-by-month table Claude generates is the most actionable part of this output.
Here is a simplified example of what the first six months might look like for a $19/month subscription business starting from zero:
| Month | Customers | Revenue | Costs | Profit |
|---|---|---|---|---|
| Month 1 | 12 | $228 | $400 | -$172 |
| Month 2 | 31 | $589 | $400 | $189 |
| Month 3 | 58 | $1,102 | $450 | $652 |
| Month 4 | 94 | $1,786 | $500 | $1,286 |
| Month 5 | 140 | $2,660 | $550 | $2,110 |
| Month 6 | 197 | $3,743 | $600 | $3,143 |
What I pay attention to most is the breakeven month and the biggest risk section.
Claude will usually flag the most dangerous assumption in your model, whether that is churn, CAC, or conversion rate, and that is exactly where I focus my early testing before I scale anything.
5. The Viral Launch Hook Generator
Getting the business model right is only half the battle. If nobody hears about your launch, none of the work from the first four prompts matters.
Prompt 5 solves that by generating 15 viral hooks for your startup across three platforms, written in the style of MrBeast’s content team.
I was skeptical of this prompt at first. It felt gimmicky. But the hooks it produces follow a proven pattern of curiosity gaps, specific numbers, and outcome-first framing that consistently outperforms generic announcement posts.
I have used outputs from this prompt directly in launch content and seen the difference in engagement.
Here is the prompt:
You are MrBeast’s hook writer applied to startup pitches. Core offer and startup: [paste your business name and one-sentence pitch]. Generate 15 viral launch hooks and titles following these rules: use MrBeast patterns including insane numbers, curiosity gaps, “I did X in Y days” and “here’s what happened when” formats. Write 5 hooks for X and Twitter, 5 for YouTube and Shorts, and 5 for TikTok, Reels, and LinkedIn. For each hook, include the first 30 seconds of spoken script in exact words. Then pick the single strongest hook and explain why it will convert.
The 15 hooks Claude generates cover a wide range of angles, which is useful because what works on LinkedIn rarely works on TikTok.
Here is a breakdown of how the hook styles differ by platform:
| Platform | Hook Style | What It Prioritizes |
|---|---|---|
| X / Twitter | Short, punchy, data-led | Shareability and replies |
| YouTube / Shorts | Curiosity gap, outcome-first | Watch time and clicks |
| TikTok / Reels | Personal story, fast hook | Retention in first 3 seconds |
| Credibility-led, results-focused | Professional trust and saves |
I always take the hook Claude picks as its strongest and test it first. If it does not perform within 48 hours, I move to the next one on the list.
Having 14 backups ready before launch removes a lot of stress from the early growth phase.
6. The 10-Slide Investor Pitch Deck Outline
Even if you never plan to raise money, building a pitch deck is one of the most clarifying exercises a founder can do.
It forces you to distill everything into a story that a stranger can understand in under ten minutes.
Prompt 6 asks Claude to act as a pitch deck specialist and produce a complete 10 to 12-slide outline using everything built in the previous prompts.
I use this output in two ways. When I need to bring on a partner or collaborator, I send them the deck outline before our first conversation.
It saves hours of back-and-forth explanation. And when I am planning content, the slide headlines become ready-made talking points for videos, newsletters, and social posts.
Here is the full prompt:
You are a pitch deck specialist who has helped raise $50M+. Using everything built so far: [paste key outputs from Prompts 1 through 5]. Create a 10 to 12 slide investor pitch deck outline. For each slide provide the slide number and title, a killer headline that states what the investor must believe, 3 to 5 bullet points of supporting content, the primary visual or proof element to include, and a one-sentence speaker note. Make every headline punchy, data-backed, and objection-proof.
The structure Claude follows maps closely to what top accelerators like Y Combinator expect from early-stage decks. Here is what a strong 10-slide outline looks like:
| Slide | Title | What It Must Prove |
|---|---|---|
| 1 | The Problem | This pain is real, widespread, and expensive |
| 2 | The Solution | Our product solves it in a way nothing else does |
| 3 | Market Size | The opportunity is large enough to matter |
| 4 | Product | Here is exactly how it works |
| 5 | Traction | People are already using and paying for this |
| 6 | Business Model | Here is how we make money and keep it |
| 7 | Competition | We know the field and have a clear edge |
| 8 | Go-to-Market | Here is how we reach our first 10,000 customers |
| 9 | Team | We are the right people to build this |
| 10 | The Ask | Here is what we need and what it unlocks |
The speaker notes Claude writes for each slide are worth paying close attention to.
They surface the objections an investor is most likely to raise at that moment in the pitch, which gives me a chance to address them before they come up.
That kind of preparation is what separates a confident pitch from one that falls apart under the first tough question.
7. The Cold Outreach and First 50 Customers Playbook
Getting your first 50 customers is the hardest part of any startup. Not because the product is not ready, but because most founders do not know how to start a conversation with a stranger that leads to a sale.
Prompt 7 fixes that by asking Claude to write a complete five-message outreach sequence and a targeting strategy for finding the right people to send it to.
I have used cold outreach sequences built from this prompt to book real conversations with potential customers before a product was even finished.
The key is that the sequence is designed to give value before it asks for anything. By the time message five arrives, the recipient already knows who I am and what I stand for.
Here is the prompt:
You are a cold outbound expert who books real meetings. Target customer profile: [paste customer profile from Prompt 1]. Create a 5-message outreach sequence for DM, LinkedIn, or email following this structure: Message 1 is a pure curiosity or value hook with no selling. Message 2 gives something free and plants a seed. Message 3 uses social proof and a mini story. Message 4 introduces soft scarcity or exclusivity. Message 5 makes the final ask with an easy yes. Also provide 5 X or LinkedIn search queries to find my first 50 perfect targets, a one-sentence follow-up cadence, and the expected reply rate and booking rate if the sequence is executed consistently.
Here is how the five-message sequence is structured and what each message is designed to accomplish:
| Message | Type | Goal |
|---|---|---|
| Message 1 | Curiosity or value hook | Start a conversation with no pressure |
| Message 2 | Free resource or insight | Build goodwill and establish credibility |
| Message 3 | Social proof and story | Show that others have seen results |
| Message 4 | Soft scarcity or exclusivity | Create a reason to act sooner |
| Message 5 | Direct ask with easy yes | Convert interest into a booked call or sale |
Claude also returns five platform search queries I can use to find my exact target customer on X or LinkedIn. That part is underrated. Instead of posting and hoping the right people find me, I use those queries to go directly to where my customers are already having conversations about the problem my product solves. That shortcut alone is worth running this prompt for.
8. The 30-Day Post-Launch Momentum Sprint
Most startups do not fail at launch. They fade out in the 30 days after it.
The initial excitement dies, the early adopters go quiet, and without a clear plan for what to do next, founders default to random activity that does not compound into anything.
Prompt 8 exists to prevent exactly that.
This prompt asks Claude to build a structured four-week growth sprint that starts the moment your first customers are in.
Each week has a different focus, specific actions, a posting and outreach cadence, KPIs to hit, and a follow-up prompt to run with Claude at the end of that week.
I treat this output as my operating plan for the first month, not a suggestion.
Here is the full prompt:
Assuming my first customers are in, build a 30-day growth sprint plan structured by week. Week 1 focuses on fixing leaks and doubling down on what is working. Week 2 introduces new traffic experiments. Week 3 builds an upsell or ascension offer. Week 4 systemizes and automates what is working. For each week provide 3 to 5 key actions, a posting and outreach cadence, KPIs to hit, and one prompt I should run with Claude at the end of that week. Close with a one-sentence projection of what hitting these targets means for revenue in the first 30 days.
Here is what the four-week structure looks like when Claude fills it out:
| Week | Focus | Primary KPI |
|---|---|---|
| Week 1 | Fix leaks, double down on winners | Churn rate and customer feedback score |
| Week 2 | New traffic experiments | New leads generated and cost per lead |
| Week 3 | Upsell and ascension offer | Revenue per customer and upgrade rate |
| Week 4 | Systemize and automate | Hours saved and repeatability of results |
The end-of-week prompts Claude suggests are one of the most practical parts of this output.
They keep me in a feedback loop with the AI throughout the sprint rather than running one big prompt at the start and then guessing my way through the rest of the month.
Each check-in prompt is designed to take the real data from that week and adjust the plan for the next one.
9. The Reality Check Audit
This is the prompt I was most tempted to skip and the one that ended up being the most valuable.
Prompt 9 asks Claude to audit your entire startup plan against what actually works in 2026 and tell you, without softening it, what you are probably getting wrong.
By this point in the sequence I am usually excited. I have a validated idea, a business model, a pitch deck, a launch plan, and a 30-day sprint.
Everything looks good on paper. That is exactly when this prompt is most needed, because excitement is the enemy of clear thinking.
Here is the full prompt:
Audit my startup plan against 2026 reality. Current summary: [paste key outputs from previous prompts]. Tell me brutally: the 5 things most founders still get wrong that I should avoid, the 3 highest-leverage moves I am probably missing, one contrarian positioning or pricing play for my niche, the biggest risk of stalling at zero to five thousand dollars in revenue, and one sentence that would multiply my results if I acted on it immediately. No fluff.
What makes this prompt different from a standard review is the contrarian positioning question. Claude does not just validate what I have already decided.
It pushes back on my assumptions about pricing, audience, and positioning in ways that often reframe the entire opportunity.
Here is the kind of output this section typically produces across different business types:
| Business Type | Common Contrarian Play Claude Surfaces |
|---|---|
| Subscription content | Charge more upfront, attract serious buyers only |
| Service business | Niche down to one industry instead of serving everyone |
| AI tool product | Give the core tool free, charge for done-for-you setup |
| Community platform | Launch with a waitlist, create scarcity before opening |
I always take the single sentence at the end of this prompt, the one Claude says would multiply my results, and write it somewhere I will see it every day during the build.
In my experience, that sentence is almost always the thing I was already avoiding because it felt uncomfortable. That discomfort is usually a signal worth paying attention to.
How to Get the Most Out of This Prompt Sequence
Running all nine prompts in one sitting is possible, but I get better results when I treat each one as a separate work session.
I run Prompt 1, sit with the output for a day, and come back to Prompt 2 with fresh eyes. That gap between sessions lets me spot things I would have glossed over if I had powered through the whole sequence in one go.
The other thing I do differently now is save every output in a single document as I go. By the time I reach Prompt 9, that document is essentially a founder’s bible for the idea.
It has the market data, the business model, the financial projections, the pitch deck, the outreach scripts, and the growth plan all in one place. That document becomes the source of truth I refer back to every time I make a decision during the build.
One thing worth saying clearly: these prompts are a starting point, not a shortcut to skipping the work. Claude will give you the framework, the numbers, and the strategy.
But the validation tests in Prompt 2 still need to be run in the real world. The outreach sequences in Prompt 7 still need to be sent to real people.
The sprint plan in Prompt 8 still needs to be executed week by week. What the prompts do is compress months of guesswork into a few focused hours, and that compression is where the real value lives.
If you have an idea sitting in the back of your mind that you keep telling yourself you will get to eventually, this is the sign to stop waiting. Run Prompt 1 today with your actual constraints and see what comes back.
The worst outcome is that Claude kills the idea before you waste six months on it. The best outcome is that you end up with a startup plan you actually believe in.
Either way, you are better off than you were before you started.
If you want to go deeper on AI tools, prompts, and building real income online, come join us at Your First 5K Club where thousands of members are doing exactly that.
