From IT support calls to building LaptopFinderAI
For years, I worked as an IT support engineer. One of the most common requests I handled wasn’t about fixing anything, it was simply:
“What laptop should I buy?”
Students, parents, small business owners, developers, creatives. Everyone had a different job to do, but they were all overwhelmed by specs, models, and marketing noise. Over time, I started noticing patterns: certain laptops consistently worked well for certain use cases, and others caused repeat complaints.
LaptopFinderAI came out of that experience.
Instead of another “best laptops” list, I wanted something that starts with how someone actually uses their laptop, then narrows options based on real constraints like budget, performance tradeoffs, battery life, and portability. The goal was to help people get to a reasonable decision quickly, with explanations, not just recommendations.
Technically, the product is built to be practical and scalable:
Next.js for performance, SEO, and credibility
OpenAI to encode structured decision logic and explain tradeoffs clearly
Supabase for auth and data without heavy backend overhead
Stripe for subscriptions and payments
One of the biggest lessons so far: structure matters. Early versions of the product didn’t perform as well because search engines and users didn’t fully trust or understand it. Rebuilding with clearer architecture and better UX has made a noticeable difference.
The product is live at laptopfinderai.com, and I’m still iterating based on real feedback.
If you’ve built a product that started as “just a recurring problem at work,” I’d love to hear how you turned that insight into something real.
Love that this came from actual IT support experience rather than "I had an idea in the shower." The fact that you already knew which laptops caused repeat complaints is a huge advantage over someone just scraping spec sheets and feeding them to GPT. The tech stack choice makes sense too. Next.js for SEO is smart here because laptop recommendation queries have real search volume, and most competing content is either affiliate blogs or generic listicles. If you can rank for long-tail stuff like "best laptop for video editing under $1000" with actually personalized results, that's a solid moat. One thing I'd think about early: how are you keeping the laptop database current? New models drop constantly and prices shift weekly. That's where a lot of recommendation tools quietly break down — the suggestions get stale and users lose trust without knowing why. Even a simple "last updated" timestamp on recommendations could help with credibility.
Congrats! Great origin story. Turning repeated real-world support questions into a structured decision tool makes a lot of sense, especially with clear tradeoff explanations instead of generic lists.
Quick security question: since you’re using OpenAI for structured recommendations, how are you validating and constraining model outputs to prevent prompt manipulation or injected content from influencing recommendations or links?
Congrats on the launch, looks solid. How are you currently thinking about acquiring early users and gathering feedback?