Skip to main content

​
Analytics

Monitor agent performance and usage across AI for Work.

​
Overview

Analytics provides visibility into:
  • Agent usage and adoption
  • Conversation quality and outcomes
  • User satisfaction metrics
  • Performance trends over time

​
Dashboard View

The dashboard provides an overview of key metrics.

​
Key Metrics

MetricDescription
Total conversationsNumber of user interactions
Active usersUnique users engaging with agents
Resolution rate% resolved without escalation
Avg. response timeTime to first agent response
User satisfactionAverage rating from feedback

​
Time Filters

View metrics for:
  • Last 24 hours
  • Last 7 days
  • Last 30 days
  • Custom date range

​
Agent Breakdown

Compare performance across agents: 
Agent Performance (Last 7 Days)
─────────────────────────────────────────────────
Agent              Conversations  Resolution  Satisfaction
─────────────────────────────────────────────────
IT Assist          1,234          87%         4.2/5
HR Assist          892            92%         4.5/5
Knowledge Bot      567            78%         4.0/5
─────────────────────────────────────────────────

​
Data View

Detailed data exploration and export.

​
Conversation Logs

Browse individual conversations:
  • Filter by agent, user, date, outcome
  • View full conversation transcript
  • See tool calls and responses
  • Review feedback and ratings

​
Export Options

Export data for analysis:
FormatContents
CSVTabular conversation data
JSONFull conversation objects
ReportFormatted PDF summary

​
Custom Queries

Build custom data views: 
-- Example: High-value conversations
SELECT
  agent_name,
  user_id,
  conversation_length,
  resolution_status
FROM conversations
WHERE
  date >= '2024-01-01'
  AND tools_used > 2
ORDER BY conversation_length DESC

​
Metrics Deep Dive

​
Usage Metrics

Track adoption and engagement:
MetricDescription
Daily active usersUnique users per day
Conversations per userAvg. conversations per user
Peak usage timesBusiest hours/days
Channel distributionUsage by Teams, Slack, Web

​
Quality Metrics

Measure conversation quality:
MetricDescription
Resolution rateSuccessfully resolved queries
Escalation rateHanded off to humans
Repeat queriesSame question asked again
Conversation lengthMessages per conversation

​
Performance Metrics

Monitor system performance:
MetricDescription
Response latencyTime to generate response
Tool success rateSuccessful tool executions
Error rateFailed responses
AvailabilitySystem uptime

​
Reports

​
Automated Reports

Schedule regular reports:
  1. Go to Analytics → Reports
  2. Click Create Report
  3. Select metrics to include
  4. Set schedule (daily, weekly, monthly)
  5. Add recipients

​
Report Templates

TemplateContents
Executive SummaryHigh-level KPIs and trends
Agent PerformanceDetailed per-agent metrics
User AdoptionUsage growth and engagement
Quality AnalysisResolution rates and feedback

​
Alerts

Configure alerts for important changes:

​
Alert Types

AlertTrigger
High escalationEscalation rate exceeds threshold
Low satisfactionUser rating drops below target
Error spikeError rate increases suddenly
Usage anomalyUnusual traffic patterns

​
Configuration

Alert: High Escalation Rate
Condition: escalation_rate > 25%
Window: 1 hour
Channels:
  - email: ops-team@company.com
  - slack: #agent-alerts

​
Best Practices

​
Monitoring

  • Review dashboard daily for anomalies
  • Set up alerts for critical thresholds
  • Track trends over time, not just snapshots
  • Compare performance across agents

​
Optimization

  • Identify high-escalation topics for improvement
  • Review low-satisfaction conversations
  • Update knowledge sources based on failed queries
  • A/B test instruction changes

​
Reporting

  • Share regular reports with stakeholders
  • Include both quantitative metrics and qualitative insights
  • Track progress against goals
  • Celebrate improvements