Score every inbound reply before a human sees it.

Most tradespeople do not lose jobs because the price was wrong. They lose jobs because the hot reply sat in the inbox for nine hours while they were on a roof, and the customer called the next guy who texted back in twelve minutes. The AI Lead Qualifier exists to fix that specific failure — the undifferentiated inbox — and to do it without inventing a funnel.

The scoring rubric lives in a trades-language prompt library, not a generic chatbot script: it reads inbound replies for service, urgency, and budget signals the way a dispatcher would read them, then routes accordingly. Hot leads go to your inbox as a single row, pre-tagged. Warmed and cold replies get filed, not deleted, so you can still quarter-end-review what you said no to and why.

It is deliberately not a chatbot. There is no back-and-forth, no script reveal, no MBA funnel jargon. The agent scores what already arrived; it does not generate new conversation. That distinction matters: scoring a real reply is a low-risk, high-signal automation; open-ended chat on a real lead is the opposite — high risk, low signal, and exactly where contractor brands go to die.

If the inbox problem is yours, install the Lead Qualifier, point it at your trade and service area, and let the first thirty replies tell you whether the rubric matches the way you actually read leads. Most contractors tweak two or three scoring thresholds in the first week — the prompts are yours to edit, not a black box.