The LED Running Shirt Story: From Validation to Pivot

The idea seemed simple enough. Create a high-tech running shirt with an LED display showing real-time stats.

I’ll admit it was ambitious. But this wasn’t about becoming a billionaire overnight. I wanted to learn how AI tools could help validate and develop a product idea from scratch.

Getting Started

I began with a bundle of AI tools from Lenny Rachitsky. The toolkit included Cursor, Lovable, Bolt, Replit, v0, Perplexity Pro, and Notion AI. I also brought in Gemini, ChatGPT, Runway, and Claude to round out my workflow.

Finding My AI Advisors

Two AI advisors became my go-to guides:

• Lennybot (LLM based on all 200+ podcast episodes and 500+ newsletters by Lenny Rachitsky)
• AIProductGPT (created by Marily Nika, my teacher in the Mavens AI PM bootcamp).

I asked them both the same question about validating my LED shirt idea. Their responses were helpful, with similar suggested steps but unique insights. I used Notion AI to combine their advice into a comprehensive plan.

Phase 1: Validation

• Interview 10–15 runners who use Garmin devices
• Document insights about displaying pace and distance
• Gather feedback on concerns like privacy, battery life, and weight
• Research price sensitivity
• Create and distribute a runner survey
• Share the survey in running communities (Reddit, Strava)
• Create visual mockups using Bolt
• Build a landing page with Lovable
• Test social media response with a pre-order button

Phase 2: Prototype Development

• Set up Garmin Connect IQ SDK for data transmission
• Configure ESP32 microcontroller and LED display panels
• Design power management and Bluetooth connectivity
• Integrate everything into shirt material
• Build the first complete prototype

Phase 3: Marketing

• Develop taglines and FAQ sections
• Create technical specifications and product roadmap
• Submit to Product Hunt and set up Kickstarter
• Integrate with Strava forums

• Interview 10–15 runners who use Garmin devices
• Document insights about displaying pace and distance
• Gather feedback on concerns like privacy, battery life, and weight
• Research price sensitivity
• Create and distribute a runner survey
• Share the survey in running communities (Reddit, Strava)
• Create visual mockups using Bolt
• Build a landing page with Lovable
• Test social media response with a pre-order button

• Set up Garmin Connect IQ SDK for data transmission
• Configure ESP32 microcontroller and LED display panels
• Design power management and Bluetooth connectivity
• Integrate everything into shirt material
• Build the first complete prototype

• Develop taglines and FAQ sections
• Create technical specifications and product roadmap
• Submit to Product Hunt and set up Kickstarter
• Integrate with Strava forums

Phase 1

Now I started with the first phase, not being certain in which phase my idea was going to end.

Creating the Survey

The validation process was entirely AI-driven. I asked both Lenny and Marily for survey questions. Then I used Notion AI to combine their suggestions into a comprehensive interview format.

I tried using Gemini’s “Help me create a form” feature, but it wasn’t available in my country. Instead, I successfully used the “AI Form Builder” add-on to create a Google Form from my compiled questions.

The Feedback Flood

The response exceeded my expectations. This looks promising!

I had Marily write posts for Reddit and Strava communities, linking to the questionnaire. Strava rejected my post. Reddit didn’t (obviously).

I posted in two Reddit communities and generated 42 thoughtful comments (Post 1 and Post 2). The reactions ranged from “Fck no” and “Stupid as fck” to “This would be AWESOME!”

I also collected 44 responses through the Google Form, surprisingly quickly.

Making Sense of It All

My analysis process had multiple steps:

• Export Google Form responses to a Notion table
• Use ChatGPT to analyze and summarize Reddit comments
• Have Notion AI combine feedback from both sources into actionable insights

The Hard Truth

The feedback analysis revealed some uncomfortable realities. That was a bummer. Apparently mu idea sucked.

About 80% of respondents said they “Definitely would not” wear the shirt. Major concerns centered around privacy—users strongly opposed publicly displaying personal metrics. Practical issues like battery life, weight, heat, and washing created significant barriers. Most users found the concept redundant with their existing devices.

This is where the original idea ended. Time to pivot.

The Silver Lining

Based on the comprehensive feedback, I identified a more promising direction:

Professional Pacer Equipment

• Clear use case for marathon and race organizations
• Professional context eliminates social anxiety concerns
• Controlled environment addresses practical concerns
• B2B model supports higher price points

What I Learned

This journey showed me something valuable. While the original consumer concept wasn’t viable, the validation process led to a more focused and promising opportunity in professional racing. Most of all it taught me about using some AI tools (in this pahse mostly Notion AI, Perplexity, Claude, ChatGPT, Runway). The negative feedback did not allow me to start experimenting with tools like Cursor, Lovable, Bolt, Replit, v0.

Now it is time to pivot and hopefully my idea goes to a later phase where I can start using those.

To be continued…

 

If you are interested to hear more. On this page is a form where you can leave your email address: https://misat.nl/great-product-ideas/led-running-shirt/