Analytics & Performance Metrics

Track your AI agent's performance, understand customer behavior, and make data-driven decisions to improve your support operations.

Accessing Analytics#

Analytics does not have a dedicated sidebar link. Access it directly at /analytics or from the main dashboard.

Overview Dashboard#

Your analytics home shows the most important metrics at a glance.

Key Performance Indicators (KPIs)#

Date Filter

Use the date selector to filter all metrics. Available options:

• Last 7 days
• Last 30 days
• Last 90 days
• Custom Range

Total Conversations

• How many customer interactions you've had
• Tracks growth over time
• Filter by date range to see trends

Resolution Rate

• Percentage of conversations solved without human intervention
• Good: 70-85%
• Excellent: 85%+
• Needs work: Below 60%

Average Response Time

• How quickly customers get their first answer
• Excellent: Under 3 seconds
• Good: 3-10 seconds
• Acceptable: 10-30 seconds

Customer Satisfaction

• Based on customer star ratings (out of 5)
• Excellent: 4.5+/5
• Good: 3.75–4.5/5
• Needs improvement: Below 3.75/5

Escalation Rate

• Percentage of conversations requiring human help
• Excellent: Under 15%
• Good: 15-30%
• High: Over 30% (consider more training)

Conversation Analytics#

Dive deeper into conversation data.

Conversation Volume#

Daily/Weekly/Monthly views:

• Line chart showing conversation volume over time
• Identify busy periods
• Plan staffing accordingly
• Spot unusual spikes or drops

By Channel:

• Which channels get most traffic?
• Web chat vs WhatsApp vs Email
• Allocate resources appropriately

By Time of Day:

• When are customers most active?
• Peak hours vs quiet periods
• Plan support coverage
• Consider time zone differences

Conversation Outcomes#

Resolved Successfully:

• AI handled completely
• Customer got their answer
• No escalation needed
• ✅ This is your goal!

Escalated to Human:

• Needed human intervention
• See reasons for escalation
• Identify training opportunities

Abandoned:

• Customer left before resolution
• May indicate frustration
• Could signal response time issues

Still Open:

• Active conversations
• Waiting for response
• In progress

Conversation Duration#

Average conversation length:

• How many messages exchanged?
• How long did it take?
• Shorter usually better (faster resolution)

Duration by Topic:

• Simple questions: 1-3 messages
• Complex issues: 5-10 messages
• Very complex: 10+ messages

Use this to:

• Identify areas needing better documentation
• Find topics that consistently take longer
• Optimize agent instructions

Agent Performance#

Track how each AI agent is performing.

Per-Agent Metrics#

For each agent, see:

• Conversations handled - Volume
• Resolution rate - Success without escalation
• Response time - Speed
• Customer satisfaction - Ratings
• Most common topics - What they handle most

Comparing Agents#

If you have multiple agents:

• Which performs best?
• Which needs improvement?
• Which handles specific topics better?
• Should you consolidate or specialize?

Agent Improvement Tracking#

Week over week:

• Is resolution rate improving?
• Are response times faster?
• Are customers more satisfied?

Track impact of changes:

• Added new documents → Resolution rate up?
• Updated instructions → Better responses?
• New playbook → Fewer escalations?

Topic & Intent Analysis#

Understand what customers are asking about.

Top Topics#

See most common conversation topics:

1. Order status (35% of conversations)
2. Shipping questions (20%)
3. Return requests (15%)
4. Product information (12%)
5. Account issues (10%)
   ...

Use this to:

• Prioritize documentation efforts
• Create targeted playbooks
• Staff appropriately for common issues
• Product/service improvements

Intent Detection#

What do customers want?

• Information seeking (most common)
• Problem solving
• Transaction/action needed
• Complaint/frustration

Trending Topics#

Rising:

• Topics increasing in frequency
• May indicate new product launch
• Or emerging issue needing attention

Falling:

• Topics decreasing
• May indicate successful resolution
• Or reduced interest

Time-Based Analytics#

Response Time Analysis#

First response time:

• How long until customer gets first reply
• Goal: Under 5 seconds for AI

Average response time:

• Throughout the conversation
• Should stay consistently fast

Resolution time:

• Total time from start to resolution
• Lower is usually better (unless complex issue)

Peak Times#

Busiest hours:

• When do most conversations happen?
• Plan team coverage
• Ensure AI is performing well during peaks

Quietest times:

• Best for maintenance
• Good for training updates
• Review and optimization time

Day of Week Patterns#

Weekday vs Weekend:

• Different volumes?
• Different types of questions?
• Adjust staffing and AI settings

Customer Satisfaction Metrics#

Rating System#

Customers can rate messages using a star rating out of 5.

Satisfaction Trends#

Track over time:

• Is satisfaction improving?
• Which responses get best ratings?
• Which get worst ratings?

Feedback Comments#

When customers leave feedback:

• Read the comments
• Identify patterns
• Address common complaints
• Learn from praise

Source Performance#

Document Usage (Planned)#

Planned: Document usage analytics are not yet available.

Planned metrics include: times referenced, success rate when used, customer satisfaction, and last used date.

Playbook Performance (Planned)#

Planned: Playbook performance analytics are not yet available.

Planned metrics include: times triggered, completion rate, resolution rate, average duration, and customer satisfaction.

Channel Performance#

By Channel Comparison#

Compare metrics across channels:

Insights:

• Which channels are most efficient?
• Where should you focus improvements?
• Are expectations different by channel?

Custom Reports#

Note: Analytics → Reports is accessible directly at /analytics/reports, not via sidebar navigation.

Creating a Custom Report#

1. Go to /analytics/reports
2. Click Create Report
3. Choose metrics to include
4. Select date range
5. Add filters (agent, channel, topic)
6. Choose visualizations (charts, tables)
7. Save and generate

Report Types#

Summary Report:

• High-level overview
• Key metrics only
• Great for executives

Detailed Analysis:

• Deep dive into specific area
• Multiple metrics and dimensions
• For operations team

Performance Report:

• Agent and playbook performance
• Identify training needs
• Quality assurance

Trends Analysis:

• Changes over time
• Seasonal patterns
• Growth tracking

Scheduling Reports#

Coming soon: Scheduled/automatic reports are not yet implemented.

Exporting Data#

Coming soon: Data export (PDF, CSV) is not yet implemented.

Interpreting Your Data#

Understanding Trends#

Improving metrics:

• ✅ What changed recently?
• ✅ What should you keep doing?
• ✅ Can you apply this to other areas?

Declining metrics:

• ❌ What changed?
• ❌ External factors? (seasonality, marketing campaign)
• ❌ What needs to be fixed?

Red Flags to Watch#

Resolution rate dropping:

• New types of questions?
• Agent needs more training?
• Documentation outdated?

Response time increasing:

• Too many conversations?
• Need to optimize?
• Technical issues?

Satisfaction decreasing:

• Read recent conversations
• What's frustrating customers?
• Quick wins to improve?

Escalation rate rising:

• Why are more going to humans?
• Can AI handle these?
• Need new playbooks?

Using Analytics to Improve#

Weekly Review Routine#

Every week, check:

1. Key metrics vs last week
2. Any unusual spikes or drops?
3. Read 5-10 recent conversations
4. Check top escalation reasons
5. Review lowest-rated responses
6. Identify one thing to improve

Monthly Deep Dive#

Every month:

1. Full trend analysis
2. Compare to previous months
3. Review all agents and playbooks
4. Identify documentation gaps
5. Plan next month's improvements
6. Set specific goals

Setting Goals#

Example goals:

• Increase resolution rate from 75% to 80%
• Reduce escalation rate by 10%
• Improve satisfaction score to 90%
• Cut response time to under 3 seconds

Track progress:

• Monitor weekly
• Adjust tactics if not improving
• Celebrate when goals reached!

Common Analytics Questions#

Why do my numbers seem low at first?#

Brand new agents need time to learn and optimize. Expect metrics to improve over 2-4 weeks as you:

• Add more documents
• Refine playbooks
• Adjust agent instructions

What's a "good" resolution rate?#

Depends on your industry and complexity, but generally:

• Simple FAQ-based support: 80-90%
• E-commerce support: 70-85%
• Technical support: 60-75%
• Complex B2B: 50-70%

Should I worry about negative feedback?#

Some negative ratings are normal. Focus on:

• Overall trends (improving or declining?)
• Patterns in negative feedback
• Opportunities for improvement

If satisfaction is 80%+, you're doing well!

How often should I check analytics?#

Daily: Quick glance at key metrics (5 min)
Weekly: Deeper review and planning (30 min)
Monthly: Comprehensive analysis (2 hours)

Can I compare my performance to others?#

Industry benchmarks coming soon! For now, compare to your own historical performance.

Next Steps#

• Use insights to improve your Agents
• Create better Playbooks based on data
• Add Documents for low-performing topics
• Monitor your Conversations for escalation patterns