AI SaaS keeps hitting the same two questions: moat and metering

Momentic (YC W24) launched agents that catch bugs by using your app. The top comments were: "So is this yet another SaaS product that's going to charge me for token usage?", "So, what's the moat?", and what it does that UAT doesn't. Kvittas, an expense-splitting app, got a reply that iOS 27 now does this natively. Bema, a work messaging tool, was asked what "connect with your current messaging provider" actually looks like and whether any company would give it full Slack access. In a related thread, Anthropic stopped letting Claude Code subscriptions run through OpenClaw because such tools "put an outsized strain on our systems," and users asked why paid-for capacity counts as strain.

The pattern: buyers are tired of variable, token-based bills, and every thin product is one platform update away from being a feature. Dependence on a model vendor's pricing is its own risk. Anthropic changed the rules mid-flight for a whole class of tools.

The fair counterpoint is that these are launch-day comments, and the best ones are just asking for specifics. Momentic's pitch is concrete enough to answer them.