After four years as Canva’s first product growth hire, I left to start a SaaS discovery and review platform.
The hardest part wasn't the product-it was getting real reviews so I build subscribed.fyi firts. Over 18 months, we manually vetted a network of 1,000+ human testers. Shortly after, buyer behavior shifted from traditional search to tools like ChatGPT, putting our platform in direct competition with a chat box.
However, we noticed something critical: AI models weren't inventing recommendations. They were reading and citing the exact reputation footprint we had spent two years building-reviews, community threads, and third-party mentions.
This led us to build Subsig. We help B2B and ai software companies earn AI recommendations by fixing the upstream signals models actually read. AI visibility is a symptom; if you're invisible in ChatGPT, your reputation footprint is not visible.
The question I'm navigating: how much of a human can you automate before losing the exact thing that makes it defensible? If you’ve built a community-driven approach and would love to hear your thoughts!
The strongest signal is the founder's accumulated distribution asset: 1,000+ vetted testers and a long-built reputation footprint. Subsig isn't starting from generic AI-visibility tooling; it appears to be built on an existing evidence base.