Stackly — AI Churn Prevention System for B2B SaaS

 | CHALLENGE
 Stackly's 900-customer SaaS platform had a monthly churn rate of 8.2% — significantly above the 3–5% industry benchmark. By the time the customer success team noticed disengagement, accounts had already decided to cancel. There was no early warning system and all re-engagement outreach was manual and inconsistent. $420,000 in ARR was at risk annually from preventable churn.
 
 SOLUTION
 FLYZEO built a churn prediction and re-engagement system that monitors 12 usage signals daily, scores each account against a risk model and automatically triggers personalised re-engagement flows — escalating to human CS reps only for highest-risk accounts.
 
 Key automations delivered:
 - Daily monitoring of 12 in-app usage signals: login frequency, feature adoption, support ticket volume, session duration
 - AI risk scoring model flags at-risk accounts 21–30 days before typical cancellation intent
 - Automated personalised email sequences triggered at each risk tier — educational content, success stories, direct outreach offers
 - Highest-risk accounts automatically escalated to CS rep in Slack with full account context and suggested action
 - Weekly churn risk dashboard delivered to CS team every Monday morning
 - Post-cancellation win-back sequence for lost accounts
 
 TECH STACK: n8n, OpenAI GPT-4o, HubSpot, Segment, Intercom, PostgreSQL, Slack
 
 TIMELINE: Deployed in 10 days
 
 RESULTS
 - Monthly churn reduced from 8.2% to 3.1% within 90 days
 - $420,000 in ARR retained in the first 6 months
 - 62% of flagged at-risk accounts successfully retained
 - CS team time on manual outreach reduced by 70% — focused on highest-value conversations
 - Win-back rate for cancelled accounts improved from 4% to 18%