For us, the first 100 users came from direct outreach, user feedback, fixing obvious problems, and asking early customers for referrals. Our expected channel failed. Instead, a few early users referred others because the product solved a specific problem.

The key question is whether the AI resolves issues or merely adds another help desk layer. For a new support operation, I’d favor an AI native platform because integrated agents, workflows, context, and handoffs can reduce friction. For established teams, Zendesk or Intercom may be better because of their mature processes, integrations, and reporting.

The main benefit of a single AI agent is easier maintenance: shared knowledge, procedures, and rules eliminate duplicate fixes. However, it must still adapt its tone and responses to each channel’s expectations. Effective handoffs matter just as much, if human agents must repeat questions, omnichannel support has failed to simplify the experience.

I think skipping the waitlist makes sense if the product is actually ready for people to use. A waitlist can create a nice launch moment, but it can also give you a list of interested people without telling you much about whether they’ll actually use or pay for the product.

A big thing with PocketBase seems to be that the early experience can be very different from the production experience. It’s easy to get an MVP running quickly, but the real test is what happens when you need more complex integrations, background jobs, scaling, backups, monitoring, or stricter operational requirements.