All Docs
FeaturesSaaS FactoryUpdated February 19, 2026

AI-Powered CRM: Customer Health Scoring & Proactive Engagement

AI-Powered CRM: Customer Health Scoring & Proactive Engagement

Available since: v1.0.4

The SaaS Factory Customer Management pillar delivers a fully autonomous CRM layer — no human sales or success team required. This page covers the core components: customer health scoring, usage analytics, lifecycle stage tracking, NPS, and proactive engagement.

Architecture Overview

The CRM system is composed of three layers:

  1. Schema layer — Extended customers table with health, engagement, NPS, activity, and lifecycle fields.
  2. CRM Router — Structured API surface for reading and updating customer CRM data.
  3. AI Agents — A health-scoring agent and a proactive engagement agent that run continuously without human intervention.

Customer Health Score

Every customer has a health_score (0–100) computed by the AI health-scoring agent. The score synthesizes:

  • Engagement score — How frequently and deeply the customer uses the product.
  • Last activity — Recency of the most recent meaningful product interaction.
  • NPS — Customer-reported satisfaction signal.
  • Lifecycle stage — Contextual weight applied based on where the customer is in their journey.
  • Support signals — Volume and sentiment of recent support interactions (fed from the support pipeline).

Scores are recomputed continuously as new activity data arrives. Historical scores are retained for trend analysis.

Score Interpretation

RangeMeaning
80–100Healthy — engaged, satisfied customer
60–79Stable — normal usage, no immediate risk
40–59Watch — reduced engagement, monitor closely
20–39At risk — intervention recommended
0–19Critical — churn highly likely

Engagement Score

The engagement_score (0–100) measures the frequency and depth of product usage independently of overall health. It is updated automatically from the platform's activity event stream.

Use the engagement score to distinguish between customers who are healthy because of recent NPS responses vs. those who are genuinely active in the product.

Lifecycle Stage

Each customer is automatically assigned one of the following lifecycle stages:

StageDescription
onboardingNew customer, within initial setup window
activeRegular, healthy engagement
at_riskDeclining health or engagement signals detected
churnedNo activity beyond the churn threshold window
advocateHigh health score + positive NPS — expansion candidate

Stage transitions are fully automated. When a customer's health score crosses a configured threshold, the lifecycle stage updates and the appropriate downstream actions (outreach, alerts, scoring model adjustments) are triggered.

NPS Tracking

Net Promoter Score responses are stored per customer in the nps field. NPS data is:

  • Incorporated into the health score model.
  • Used to identify advocate-stage customers eligible for referral or expansion programs.
  • Available via the CRM router for reporting and segmentation.

CRM Router

The CRM router exposes structured access to customer lifecycle data. Key capabilities:

  • Retrieve a customer's current health score, engagement score, NPS, and lifecycle stage.
  • Query customers by lifecycle stage (e.g. fetch all at_risk customers).
  • Access health score history for a given customer.
  • Trigger a manual health score recalculation for a specific customer.

Proactive Engagement Agent

The proactive engagement agent monitors the customer population continuously. When a customer's health score or lifecycle stage crosses a configured threshold, the agent automatically triggers one or more engagement actions:

  • Automated check-in — Sends a contextual message to the customer surfacing relevant resources or offers.
  • Success resource delivery — Delivers targeted documentation, tutorials, or feature highlights based on the customer's usage patterns.
  • Escalation flag — Marks high-value at-risk accounts for priority handling in the operations dashboard.

The agent operates 24/7 with no human intervention. It respects customer communication opt-out preferences at all times.

Schema Reference

All CRM fields live on the customers table:

// customers table (billing-schema.ts)
{
  // ... existing fields ...
  health_score:      number,   // 0–100, AI-computed
  engagement_score:  number,   // 0–100, activity-derived
  last_activity:     Date,     // most recent activity event timestamp
  nps:               number,   // -100–100, latest NPS response
  lifecycle_stage:   'onboarding' | 'active' | 'at_risk' | 'churned' | 'advocate'
}

Setup & Configuration

  1. Run migrations — Apply the schema migration to add the new customers columns before deploying v1.0.4.
  2. Health scoring — Enabled by default. The agent will score all existing customers on first run using available historical data.
  3. Proactive engagement — Enabled by default. Configure thresholds and outreach templates in the Customer Management settings panel.
  4. Opt-outs — The engagement agent checks communication opt-out flags automatically. No additional configuration required to respect existing preferences.

Related