What it does
- Pulls estimates straight from your CRM (HCP, ServiceTitan, etc.) so this is one source of truth.
- Visualises the pipeline by stage: draft, sent, approved, converted.
- Flags stale deals (no movement in N days) for follow-up.
- Tracks age, value, options, and the rep who wrote it.
- AI-suggested follow-ups for any estimate that's been sitting too long.
How it works
The system polls your CRM on a schedule and caches estimates in a local store so the UI stays fast even when the CRM API is slow.
Each estimate carries its history (when sent, last touched, customer activity) and a confidence score on whether it's likely to close.
When an estimate goes stale, the system drafts a follow-up message in your voice and queues it for one-click send.
The stack
HousecallPro API (extensible to any CRM with a REST API). Python polling workers. Flask + JSON cache. Best-in-class LLMs (Claude, GPT-4) for draft follow-ups.
Who it's for
Service businesses where estimates sit at the top of the sales funnel and where follow-up is the difference between $XXk and $XXXk months. Electricians, contractors, MSPs, B2B services.