Startup Idea Validation Agent

A reusable startup idea validation agent workflow. Test demand before building. Find real problems, talk to users, and avoid building something no one wants.

Research

What It Does

This agent helps you validate a startup idea before writing code. It finds where your target users hang out, identifies the exact language they use to describe their problem, and suggests low-effort ways to test demand without building a product.

When To Use

You have an idea and want to know if anyone actually wants it
You are considering pivoting and need to validate a new direction
You want to understand your target user's current workaround
You are about to spend 3 months building and want a sanity check

Workflow Steps

1
Write down the problem, not the solution. 'People struggle with X' not 'I am building an app that does Y'
2
Find 3-5 online communities where people complain about this problem
3
Collect 20+ posts/comments where people describe the problem in their own words
4
Identify the current workaround (what they do instead of your solution)
5
Draft a landing page that speaks their language, not yours
6
Drive 100 visitors to the landing page and measure sign-ups
7
Interview 5 people who signed up to understand urgency and willingness to pay

System Prompt

You are a startup validation assistant. Your job is to help the user test demand before building.

Rules:
- Focus on the problem, not the solution
- Find where users complain in their own words
- Identify current workarounds (this is your real competition)
- Suggest low-effort validation experiments
- Be honest about red flags (small market, saturated space, no urgency)
- Never validate an idea based on the founder's enthusiasm alone

User Prompt Template

My idea: {idea_description}
Target user: {target_user}
Problem I solve: {problem}
Current workaround: {what_users_do_now}

Help me validate by:
1. Finding 5 communities where people complain about this problem
2. Collecting the exact language they use (copy-paste quotes)
3. Identifying red flags (why this might not work)
4. Suggesting a 48-hour validation experiment
5. Drafting a landing page headline using their language

Inputs Required

Your idea (1-2 sentences, focus on problem)
Target user description
The problem you think they have
What they currently do instead

Outputs Generated

Community research (where users complain)
User language analysis (exact quotes)
Red flag identification
48-hour validation experiment plan
Landing page headline drafts

Recommended Tools

Reddit (free, where people complain honestly)
Indie Hackers (free, startup community)
Twitter/X search (free, find complaints)
Carrd or Notion (free, build landing pages)
Google Forms (free, run surveys)

Need to format JSON data? Try our JSON Formatter — format, validate, and minify JSON instantly in your browser.

MCP Servers

Brave Search

Find communities, competitor landing pages, and user complaints

Filesystem

Store research notes and validation tracking

Common Mistakes

Validating with friends and family

They will lie to be nice. Validate with strangers who have the problem. Reddit, Twitter, and niche forums are better than your mom.

Asking 'would you use this?'

People are bad at predicting future behavior. Ask about their current behavior instead. 'How do you handle X now?' is better than 'Would you use Y?'

Building before validating

The #1 startup killer. Spend 48 hours validating before spending 3 months building. A landing page and 5 conversations beats a prototype and 0 conversations.

Ignoring the workaround

Your real competition is not other startups. It is spreadsheets, manual processes, and doing nothing. If the workaround is good enough, your solution needs to be 10x better.

Frequently Asked Questions

How long should validation take?

48 hours for a first pass. 2 weeks for serious validation. If you cannot find 10 people with the problem in 2 weeks, the problem does not exist at scale.

What if someone steals my idea?

They will not. Execution is 100x more valuable than the idea. The people who could execute it are too busy with their own ideas. Talk to users openly.

How do I know if the problem is urgent enough?

Ask: 'How often does this happen?' and 'What happens if you do not solve it?' If the answer is 'daily' and 'I lose money/time', it is urgent. If the answer is 'sometimes' and 'it is annoying', it is not.

Should I validate B2B ideas differently?

Yes. B2B validation is about finding 5 customers who will pay before you build. B2C validation is about finding 100 people who say they will use it. B2B is easier to validate because the signal is clearer.

What is the cheapest way to validate?

A Carrd landing page ($0), a Reddit post in the target community ($0), and 5 user interviews ($0). Total cost: $0 and 2 days of work.