Dashboard Metrics

Learn what each metric means and how to use them to improve your support.

Conversation Metrics

Total Conversations

The total number of chat conversations started.

Includes:

• All customer-initiated chats
• Proactive message responses
• Both AI-only and human-handled

Use this to:

• Track overall volume
• Plan staffing needs
• Measure marketing impact

New vs Returning

Breakdown of visitor types:

• New — First-time visitors
• Returning — Have chatted before

Use this to:

• Understand customer loyalty
• Identify repeat issues
• Measure retention

AI Performance Metrics

AI Resolution Rate

Percentage of conversations resolved by AI without human intervention.

Calculation:

AI-only resolutions / Total conversations × 100

Healthy range: 60-80%

If too low:

• Add more knowledge base content
• Review low-confidence topics
• Improve content quality

If too high:

• Consider if customers are leaving unsatisfied
• Review for false positives

Average Confidence

Average AI confidence score across all responses.

Healthy range: 70-85%

Low confidence indicates:

• Knowledge gaps
• Unclear customer questions
• Need for more training data

Escalation Rate

Percentage of conversations escalated to humans.

Includes:

• AI-initiated escalations
• Customer-requested escalations

Target: 20-40% depending on complexity

Escalation Reasons

Breakdown of why conversations escalated:

Response Metrics

Average Response Time

How quickly responses are sent.

AI Response Time:

• Should be instant (< 2 seconds)
• Measures time from message to response

Human Response Time:

• Time from escalation to first human reply
• Target: < 5 minutes during business hours

First Response Time

Time to first response (AI or human).

Why it matters:

• Impacts customer satisfaction
• Faster = better experience
• Goal: < 30 seconds

Customer Satisfaction

CSAT Score

Customer satisfaction rating (1-5 stars).

Collection:

• Shown after conversation ends
• Optional for customers
• Higher is better

Benchmarks:

• Below 3.5: Needs improvement
• 3.5-4.0: Average
• 4.0-4.5: Good
• Above 4.5: Excellent

Response Rate

Percentage of customers who rate:

Low response rate?

• Make rating more visible
• Reduce rating friction
• Time request appropriately

Team Metrics

Conversations Per Agent

How many conversations each agent handles.

Use this to:

• Balance workload
• Identify capacity issues
• Plan schedules

Agent Response Time

Average response time per team member.

Compare to identify:

• Top performers
• Training needs
• Workflow issues

Agent Resolution Rate

Percentage each agent successfully resolves.

Factors affecting this:

• Conversation complexity
• Product knowledge
• Tools and resources

Trend Analysis

Week-over-Week

Compare current week to previous:

• Volume changes
• Rate improvements
• Seasonal patterns

Month-over-Month

Longer-term trends:

• Growth patterns
• Improvement trajectories
• Anomaly detection

Knowledge Gaps

Low Confidence Topics

Questions the AI struggles with:

• Listed by frequency
• Sorted by confidence
• Links to add content

Unmatched Queries

Questions with no knowledge base match:

• Completely new topics
• Variations of existing topics
• Opportunities for content

Using Metrics

Daily Review

Check daily:

• Total conversations
• Escalation rate
• Response times
• Any alerts

Weekly Analysis

Review weekly:

• AI resolution rate trends
• Team performance
• CSAT scores
• Knowledge gaps

Monthly Planning

Use for planning:

• Staff scheduling
• Content priorities
• Training needs
• Tool improvements

Best Practices

Set Benchmarks

Establish baseline metrics:

• Measure for 2-4 weeks
• Set realistic targets
• Track improvement

Regular Reviews

Schedule analytics reviews:

• Daily quick check
• Weekly team review
• Monthly deep dive

Take Action

Metrics without action are useless:

• Identify issues
• Plan improvements
• Measure impact

Next Steps

• Export reports
• Improve AI training
• Adjust confidence settings