Signals

Signals help you monitor specific topics in contact conversations. Define signals for things you want to track—like feature requests, pricing questions, churn indicators, or product feedback—and the AI automatically applies them.

Accessing Signals

Navigate to Training > Signals in the sidebar to manage your signals.

How Signals Work

1. You define signals with descriptive names and definitions
2. When contacts chat, AI analyzes the conversation
3. Relevant signals are automatically applied
4. You view signal distribution in the Insights dashboard

Signals give you visibility into what contacts are talking about without reading every conversation.

Creating a Signal

1. Navigate to Training > Signals
2. Click Add Signal
3. Enter a Signal Name (e.g., "Feature Request", "Pricing Question")
4. Enter a Description explaining what this signal represents
5. Toggle Active on (AI will look for this signal)
6. Click Create Signal

Writing Good Descriptions

The description helps the AI accurately identify when a tag applies. Be specific:

Good descriptions:

"Contact is asking about features we don't currently offer or suggesting improvements to existing features."

"Contact expresses frustration, disappointment, or indicates they may stop using the product."

"Contact is asking about pricing, costs, plans, or billing."

Poor descriptions:

"Feature stuff" (too vague)

"Questions" (not specific enough)

Managing Tags

Active vs Paused

• Active - AI will analyze conversations for this signal
• Paused - AI won't apply this signal (historical data preserved)

Toggle status using the Pause/Activate button on each signal card.

Editing Signals

1. Click Edit on a signal card
2. Update the name or description
3. Click Save Changes

Changes take effect immediately for new conversations.

Deleting Signals

1. Click Delete on a signal card
2. Confirm deletion

Viewing Signal Analytics

Signal data appears in the Insights Dashboard:

Signal Distribution

A bar chart showing which signals are applied most frequently:

• Click any bar to view conversations with that signal
• Trend indicators show if signals are increasing or decreasing
• Filter by date range to see patterns over time

Donut Chart

Visual breakdown of signal share across your conversations.

Alerts Panel

When signals trend up or down significantly, alerts appear:

• "Feature Request trending up +23%"
• "Churn Risk signals increased this week"

Example Signals

Product Feedback

Sales Intelligence

Support Topics

Best Practices

Start with Key Signals

Begin with 5-10 high-value signals rather than trying to track everything:

• What questions matter most to your business?
• What topics indicate sales opportunities?
• What signals warn of contact issues?

Review and Refine

Monitor how signals are being applied:

• Are relevant conversations being tagged with signals?
• Are irrelevant conversations getting false positives?
• Refine descriptions based on results

Use Insights

Turn signal data into action:

• Rising "Churn Risk" → Proactive outreach
• Many "Feature Requests" for same thing → Product roadmap input
• Frequent "Pricing Questions" → Improve pricing page clarity

Coordinate with Team

Share signal insights:

• Product: Feature requests and feedback
• Sales: Upgrade interest and competitor mentions
• Support: Common issue patterns
• Marketing: Contact sentiment trends

Integration with Rundowns

Signal tag data is included in AI-generated Rundowns:

• Which signals appeared most often
• Trends compared to previous periods
• Suggested actions based on patterns

Learn more about Rundowns →