Signals
Signals
Automated detection of important events and patterns across your customer accounts.
Overview
Signals are automatically detected events or patterns that indicate changes in customer health, risk, or opportunity. Karolina processes data from all your integrations to surface actionable insights.
Signal Types
Risk Signals
| Signal | Description | Severity |
|---|---|---|
| Usage Drop | Significant decline in product usage | High |
| Support Spike | Increase in support tickets | Medium |
| Sentiment Decline | Negative trend in communication | High |
| Champion Left | Departure of key stakeholder | Critical |
| Competitor Mention | References to competitor products | Medium |
| Payment Issue | Late or failed payments | Critical |
| Low Engagement | Decreased email/meeting activity | Medium |
| Contract Negotiation | Hints of competitive evaluation | High |
Opportunity Signals
| Signal | Description | Potential |
|---|---|---|
| Usage Threshold | Approaching plan limits | High |
| High Adoption | Heavy feature usage | Medium |
| Positive Sentiment | Strong positive communications | Medium |
| Executive Interest | Increased C-suite engagement | High |
| Expansion Ready | Growth indicators | High |
| Feature Request | Requests for advanced features | Medium |
Health Signals
| Signal | Description |
|---|---|
| Health Drop | Sudden decrease in health score |
| Health Recovery | Improvement after at-risk period |
| Risk Resolved | Previously flagged risk no longer present |
Signal Sources
Signals are generated from:
| Source | Data Processed |
|---|---|
| Product Analytics | Usage events, feature flags |
| Message sentiment, response times | |
| Calendar | Meeting patterns, attendee changes |
| Support | Ticket volume, sentiment, topics |
| CRM | Account changes, opportunity updates |
| Stripe | Payment behavior, subscription changes |
| Slack | Team communications |
Signal Properties
Each signal has:
| Property | Description |
|---|---|
| Type | Category of signal |
| Severity | Impact level (Critical, High, Medium, Low) |
| Confidence | AI confidence in detection (0-100%) |
| Account | Which account triggered it |
| Detected At | When it was identified |
| Factors | What contributed to this signal |
Viewing Signals
Signal Feed
- Go to Signals in the sidebar
- See all recent signals across accounts
- Filter by:
- Type (Risk, Opportunity, Health)
- Severity
- Account
- Date range
- Sort by:
- Newest first
- Severity
- Confidence
Account Signals
View signals for a specific account:
- Open an account
- Go to Signals tab
- See all signals for that account
- Click any signal for details
Signal Details
Click a signal to see:
Summary
Quick overview of what was detected.
Contributing Factors
The specific data points that triggered this signal:
Example: Usage Drop
- Login frequency down 60% vs 30-day average
- API calls down 45%
- Active users decreased from 25 to 12
Recommended Actions
AI-suggested next steps:
1. Reach out to champion to understand原因
2. Schedule an executive check-in
3. Review recent product changes
Related Signals
Other signals that may be connected:
Related: Support Spike (2 days ago)
Related: Sentiment Decline (1 week ago)
Taking Action on Signals
Acknowledge
Mark a signal as seen:
- Click the signal
- Click Acknowledge
- Optionally add a note
Snooze
Temporarily hide a signal:
- Click the signal
- Click Snooze
- Select duration:
- 1 hour
- 1 day
- 1 week
- Custom
Create Action
Turn a signal into an actionable task:
- Click the signal
- Click Create Action
- Fill in details:
- Title
- Due date
- Assignee
- Notes
- Click Create
Add to QBR
Include in quarterly business review:
- Click the signal
- Click Add to QBR
- Select or create QBR deck
Signal Configuration
Alert Preferences
Set up notifications for new signals:
- Go to Settings → Notifications
- Click Signal Alerts
- Configure:
- New critical signals: Immediate
- New risk signals: Daily digest
- New opportunity signals: Weekly digest
Signal Thresholds
Adjust when signals are triggered:
- Go to Settings → Signals
- Configure thresholds:
- Usage drop % before alert
- Sentiment threshold
- Support ticket volume triggers
- Save changes
Suppressions
Prevent unwanted signals:
- Go to Settings → Signals
- Click Suppressions
- Add rules:
- Suppress [signal type] for [account]
- Suppress [signal type] until [date]
Signal Intelligence
Karolina uses AI to:
Correlate Signals
Connect related signals across time:
Signal Chain:
Usage Drop (Mar 15) → Support Spike (Mar 17) →
Champion Left (Mar 19) → Churn Risk (Mar 21)
Predict Outcomes
Estimate likely account trajectories:
Risk Assessment:
65% probability of churn in next 60 days
Key factors: Usage drop, executive departure
Suggest Interventions
Recommend specific actions:
Recommended Actions:
1. Immediate: Schedule check-in with new champion
2. This week: Review product adoption gaps
3. Next week: Propose refresher training
Best Practices
Daily Review
- Check Critical signals first
- Review new Risk signals
- Look for signal patterns
Signal Investigation
When a signal appears:
- Read the contributing factors
- Check the account timeline
- Look for related signals
- Review recent changes
- Take appropriate action
Preventing Noise
- Configure thresholds appropriately
- Suppress known non-issues
- Focus on high-confidence signals
- Combine signals with context
Troubleshooting
| Issue | Solution |
|---|---|
| Too many signals | Adjust thresholds, increase confidence filter |
| Missing expected signal | Check integration data, verify signal enabled |
| False positives | Mark as suppressed, add context notes |
| Signal not updating | Check for sync issues with integrations |