Customer Research
Overview
Customer Research transforms raw CRM data into actionable intelligence on every customer in your database. Instead of manually digging through job history to understand a customer's habits, the AI does the analysis automatically and presents a structured briefing on each contact record.
The goal is practical: know which customers are at risk of leaving before they leave, know which ones are ready to buy more before you send a quote, and know which ones need a personal touch before you send an automated message.
Features
- Spend Pattern Analysis — Average spend per visit, total lifetime value, spend trajectory (growing/stable/declining) over trailing 12 months
- Service History Summary — Most frequent service types, average job complexity, preferred scheduling windows, most-used technician
- Churn Risk Score — 0–100 score predicting likelihood of customer inactivity within 90 days; based on recency, frequency, and engagement signals
- Upsell Opportunities — AI-identified services the customer hasn't tried but is likely to need based on property type, service history, and peer patterns
- Interaction Quality — Flagged patterns: late payments, complaint history, review behavior, response rate to communications
- Daily Refresh — All scores and summaries recalculate overnight; morning briefing available by 6am local time
Churn Risk Score Bands
| Score | Band | Recommended Action |
|---|---|---|
| 80–100 | Stable / Loyal | Maintain contact cadence |
| 60–79 | Moderate | Check in proactively |
| 40–59 | Elevated | Launch win-back sequence |
| 0–39 | High Risk | Personal outreach from owner |
CRUD Reference
| Object | Create | Read | Update | Delete |
|---|---|---|---|---|
| Customer Research Profile | — | ✅ | — | — |
| Churn Risk Score | — | ✅ | — | — |
| Upsell Opportunity | — | ✅ | ✅ (dismiss) | — |
| Spend Analysis | — | ✅ | — | — |
| Research Refresh Log | — | ✅ | — | — |
Notes
Customer Research is an optional module — see Plugins & Modules to enable it. Data is derived solely from CRM records; no external data sources are used. Churn risk scores are predictive, not deterministic — treat them as a prioritization signal, not a guarantee. Research profiles are visible to Admins and Dispatchers; Technician and Office roles see a simplified view on the mobile pre-visit briefing only.