
Me then: vibe code our own internal tool from scratch. Me now: keep and connect to the SaaS, vibe code the one feature they wanted $40/seat for.
6 months ago, I vibe coded an internal tool that everyone uses (likely saving us ~$1,000/mo), and I now HATE it. It’s called Voice of Customer (VoC) and pulls data from a ton of disparate endpoints to give us an idea of what customers want and how they’re feeling. At first, I felt like a genius. As a non-developer, bringing something to life from scratch was magic. The initial pass took me less than a day and the responses from VoC were ~80% accurate. But then... people started using it. They expected responses to be 100% accurate. Which led me to scramble in the back end, fixing taxonomies and other data concepts I barely understood, with Claude Code. I’ve spent a total of 5 all nighters, plus what feels like constant maintenance. That $1,000/mo savings? Poof. Unless you value my time at $0. Now, every time I get a notification that someone is trying to use VoC, I feel a pit in my stomach and know that if it doesn’t work or isn’t accurate, I’m on the hook. So yeah, go ahead and vibe code your internal tools. It's a lot of fun, especially as an experiment. But then come talk to me in 6 months. My prediction: this is a pendulum. Folks will increasingly vibe code everything (cause it's cool and fun and feels like saving money), feel the maintenance, then swing back to buying SaaS. Maybe SaaS needs a rebrand. Maintenance as a service? Security as a service? Support as a service? All of the above.









