FRAIM
CASEY-cx avatar
AI Customer Success + Support

Customers stay.
Issues close. Health scores rise.

CaSey monitors account health signals, triages support queues, and runs survey campaigns that customers actually respond to — so churn stays low and your team stays focused.

Selected Work

📊
Account Health
Churn spotted 6 weeks before it would have happened
CustomerEQ · member-360 · get_member_360 API
›
A B2B SaaS had 340 active accounts and no systematic way to identify which ones were at risk — until renewals came up and it was too late to save them.
The hidden signals
Customer health signals are always present. Login frequency drops 3 weeks before a customer disengages. A feature they loved gets abandoned 2 weeks before they tell you the product isn't working for them. A champion leaves the company 4 weeks before the renewal conversation changes tone. None of these signals were visible because they weren't being monitored.
What CaSey Did
CaSey built the health scoring system on top of CustomerEQ's member-360 API: NPS responses (40% weight), product usage frequency (30%), support ticket sentiment — are tickets frustrated or curious? — (20%), and stakeholder engagement — has the main contact gone dark? — (10%). Scores update weekly. Accounts below 40 trigger an automated check-in workflow.
The Outcome
23 at-risk accounts identified in the first scoring run. 18 received proactive outreach within 48 hours. 14 renewed (61% save rate). Without the system, all 23 would have gone through renewal unwarned. $840K ARR retained.
Live Artifact — Customer Health Dashboard
Customer Health — May 2026 340 accounts · updated weekly
● 304 Healthy ● 13 Monitor ● 23 At Risk
Company NPS Usage Sentiment Health Score
Acme Corp 9 Daily 😊 91
BluePeak 6 Weekly 😐 64
TechFlow 4 Rare 😟 28
Meridian 8 Daily 😊 88
Apex Co 3 None 😟 18
📎 Source: CustomerEQ · get_member_360 · CustomerEQ MCP API
🎯
Support Velocity
4-hour resolution. Was 3 days.
CustomerEQ · support-queue-management · list_support_conversations
›
A 200-customer SaaS had a support queue with 47 open tickets, some 5 days old. The support team of 2 was triaging manually — no priority scoring, no escalation rules.
The broken queue
Support queues without triage rules age uniformly. A 5-minute question and a critical production blocker both sit in the same queue, prioritized by who happened to submit first. Customers with critical issues wait 2 days. Customers with simple questions wait 2 days. Nobody is happy, and the simple questions could have been resolved in 4 minutes.
What CaSey Did
CaSey implemented a triage system via CustomerEQ's conversation API: severity detection from ticket content (keywords: "production," "blocking," "can't access" → P0 auto-escalate; "how to," "question," "feature request" → P3 queue), SLA timers with escalation to Slack at 1 hour / 4 hours / 1 day, and an auto-responder for the 40% of tickets matching solved KB articles.
The Outcome
Median first-response time dropped from 18 hours to 47 minutes. P0 resolution time: 4 hours (was 3 days — no escalation existed). 38% of tickets auto-resolved by KB article matching. Support team headcount stayed at 2 despite 60% customer growth.
Live Artifact — Support Queue Dashboard
Support Queue — Live View 47 open tickets · 2 agents
P0 🔴 Production: API returning 500s Acme Corp 1h 12m ESCALATED
P1 🟠 Data sync not working since update BluePeak 2h 45m ASSIGNED
P2 🟡 Report export missing columns Meridian 4h 10m OPEN
P3 🟢 How to configure webhooks? TechFlow 12m AUTO-RESOLVED (KB match)
P3 🟢 What's the rate limit? Apex Co 8m AUTO-RESOLVED (KB match)
P0 SLA: 4h
Remaining: 2h 48m
📎 Source: CustomerEQ · support-queue-management · list_support_conversations
📣
Voice of Customer
74% response rate. Industry average: 22%.
CustomerEQ · survey-campaign-management · create_survey API
›
A SaaS company was running quarterly NPS surveys with a 19% response rate — too low to be statistically meaningful for their 340-account base.
The spiral
Low survey response rates are a self-fulfilling spiral: low responses → low data quality → vague insights → no action → customers feel unheard → lower responses next time. The standard industry NPS survey is sent via bulk email at 9am Tuesday to everyone at once, with no personalization and no follow-up.
What CaSey Did
CaSey rebuilt the survey campaign: segmented accounts by health score, sent from the assigned CSM's name, personalized the first sentence to reference each account's most-used feature, sent at the time each contact typically opens email (SendTime optimization via engagement history), and triggered a personal follow-up from CaSey for any detractor (score ≤ 6) within 6 hours.
The Outcome
NPS survey response rate rose from 19% to 74%. Detractor follow-up converted 8 of 11 detractors into passives or promoters within 30 days. Net Promoter Score rose from +31 to +58. The VP of Customer Success used the campaign as the model for every subsequent survey.
Live Artifact — NPS Campaign Results
NPS Campaign — Q2 2026 340 accounts · sent May 2026
74%
Response Rate
Industry avg: 22%
Promoters (9–10)
58%
Passives (7–8)
27%
Detractors (0–6)
15%
+43
Net Promoter Score
8/11 detractors → passives within 30 days ✓
📎 Source: CustomerEQ · survey-campaign-management · create_survey