The Ultimate Niche: I Built an AI CAD Tool So My Grandpa Could Design Again
We always hear about solving problems you care about. For me, that meant solving my Grandpa's problem.
My grandfather, a lifelong maker, lost his mobility and couldn't access his workshop. We got him a 3D printer, but traditional CAD software (Fusion 360, etc.) was too complex, too many functions, tough mouse control, and a language barrier (he speaks Slovenian). He was stuck just printing other people's designs.
The Indie Hackers Lesson: The problem wasn't the printer; it was the interface.
The "Aha" Moment: He loved using ChatGPT for simple questions. That’s when the lightbulb went off: What if creating a 3D model was as simple as describing it conversationally?
I started building GrandpaCAD. We use LLMs to translate simple, natural language ("Make me a small box with a lid") directly into geometric code.
The Takeaway:
Niche Down to One User: We focused only on simple, functional parts. Exactly what he needed. Don't try to beat Fusion 360; focus where a simple conversational interface shines brightest.
Solve a Deeply Personal Pain: Seeing him successfully design and print his own creation for the first time was the true "validation." The frustration vanished, replaced by creative joy.
Features For Him: He didn't grew up using computer like we did, so icons don't mean much to him. Small buttons are hard to click,... So there is a specific feature for him "Grandpa Mode" that makes interface even simpler (for the cost of aesthetics)
If you’re struggling to find your product-market fit, try solving the most specific, personal problem you can find. It might just lead to a product that helps more people than you ever imagined.
Read the full story here: https://grandpacad.com/blog/story
Congrats on the launch, looks solid. How are you currently thinking about acquiring early users and gathering feedback?
Thanks Ajayi. Not 100% sure yet. Gonna start testing multiple channels very soon and see what works. Do you have any ideas?
That’s a good place to be, testing multiple channels early helps you find the signal faster.
If I were you, I wouldn’t test everything at once, though. I’d pick 1–2 channels where your ideal users are already actively talking about the problem you solve.
For early-stage products, I’ve seen strong results from:
• Niche Reddit communities
• Founder/indie communities
• Direct problem-based conversations instead of promotion
It’s less about volume and more about intent.
Who exactly is your ideal first user? If we narrow that down, I can suggest a more focused starting point instead of just “testing everything.”
Hobby makers, home fixers, people that want to do quick simple 3D designs but don't want to bother learning CAD software. Some statistics show that 70% of the 3D printer owners don't know how to model and a simple subset of those users could already be covered with my product.
I was thinking about Reddit, but it's so anti AI in 3D printing subreddits it's sad.