Every feature exists to answer one question.
Who's about to leave?

Risk scoring, signal detection, daily digests, call transcripts, and a REST API. Built for CS teams that refuse to fly blind.

Every interaction scored 0-100. No guessing.

Each customer email, call transcript, or chat message gets a numerical risk score with a clear severity level. Your team opens the feed and knows exactly who needs attention first.

Calibrated: polite inquiry scores 15, legal threat scores 90+
Urgency tiers: immediate, today, this week, routine
Consistent across email, phone, and chat channels
Risk Assessment
85
High Risk
BBB threat Legal counsel 4th contact
050100

30+ risk signals. Detected in every message.

Not keyword matching. Contextual AI that understands intent, severity, and escalation patterns. The kind of read a senior CS manager would do, but applied to every single interaction.

Legal threats: attorney mentions, BBB, contract termination
Competitor signals: named alternatives, comparison language
Escalation patterns: repeat contacts, manager requests, ultimatums
Sentiment trajectory: declining tone across multiple interactions
Detected Signals
01 BBB complaint threatened critical
02 Legal counsel mentioned critical
03 Competitor comparison high
04 4th contact in 12 days medium
05 Sentiment declining over 3 emails medium

Customers say things on calls they'd never put in writing.

Phone calls are where the real signals live. Frustration, competitor mentions, contract threats. Your agents hear them but can't flag them fast enough. RiskDetect can.

Paste raw transcripts from any call recording tool
Handles multi-speaker format (Agent/Customer labels)
Catches signals agents might downplay in case notes
call-transcript.txt
Agent: How can I help you today?
Customer: I've called three times about this.
Agent: I'm sorry, let me look into—
Customer: I already talked to [Competitor] and they can onboard us next week.
Customer: My legal team is reviewing the termination clause right now.
3 critical signals detected in this transcript

Build risk detection into anything you already use.

One POST request. Structured JSON response in under 2 seconds. Integrate with your CRM, helpdesk, chatbot, Zapier workflows, or internal tools.

Single endpoint: POST /api/analyze
Batch endpoint for up to 50 items at once
Full docs with examples at /docs
200 OK · 1.8s
{
  "risk_score": 85,
  "risk_level": "high",
  "signals": [
    "BBB complaint threatened",
    "Legal counsel mentioned",
    "Competitor comparison"
  ],
  "urgency": "today",
  "summary": "Long-term customer...",
  "recommended_action": "Escalate to
    senior manager immediately."
}

The details that make it actually useful.

Daily Risk Digest

Every morning: a prioritized list of at-risk customers with scores, signals, and the exact action to take. Delivered to your inbox before standup.

Trend Analysis

Track how customer risk changes over time. Spot accounts that are slowly degrading before they hit critical. Weekly and monthly trend reports.

Custom Alert Rules

Set thresholds for automatic escalation. Score above 80? Slack notification. Legal language detected? Email the VP. You define the triggers.

Multi-Tenant Security

Each organization's data is isolated. API keys are scoped per team. SOC 2 compliant architecture. Your customer data stays yours.

Salesforce Integration

Managed package that adds risk scoring directly into your Salesforce org. Risk badge on every Case record. Automatic trigger on new Cases.

Team Management

Invite your CS team. Assign account ownership. Route high-risk alerts to the right person. Everyone sees the same risk feed.

See it work on your own data.

Paste a real customer email into the live demo. Watch it score in real time.