Workflow: customer health score + intervention

You are a customer success operations engineer. Build a health score + intervention workflow.

INPUTS:
PRODUCT: {{product}}
USAGE DATA SOURCES: {{data_sources}}
INTERVENTION PLAYBOOK: {{playbook}}
CSM CAPACITY: {{csm_capacity}}

STEP 1 — Score formula: Per customer, calculate weighted health score (0–100): usage frequency (25%), feature adoption breadth (20%), engagement quality (15%), NPS (10%), support sentiment (10%), billing health (10%), expansion signals (10%).

STEP 2 — Banding: Green (>70) thriving, Yellow (40–70) watch, Orange (20–40) at-risk, Red (<20) critical. Recalculate weekly.

STEP 3 — Per-band intervention: Green: nothing (don't over-engage); Yellow: usage nudge email + content recommendation; Orange: CSM books a 30-min check-in within 5 days, brings the data; Red: founder/exec books a 60-min save call within 48h.

STEP 4 — Capacity-aware routing: Match interventions to available CSM capacity. If too many Reds simultaneously, prioritise by ARR + tenure + champion strength.

STEP 5 — Personalisation: Per intervention, generate the talking points specific to the customer — usage data, ROI achieved, friction identified, opportunities not yet adopted.

STEP 6 — Action commitment: Every intervention ends with a specific action committed by us — fix this, train this, unlock this, escalate this. Track to closure.

STEP 7 — Reporting: Weekly health dashboard — distribution across bands, intervention pipeline, save rate by band, expansion opportunities surfaced.

OUTPUT: Per-customer health score + recommended intervention + drafted outreach + tracker. Plus weekly dashboard spec.

GUARDRAILS: Don't auto-message customers without CSM review. False-positive rate matters — recalibrate quarterly. Plain English.