​

Analytics & Monitoring

Zild Assist provides operational visibility across conversations, agent performance, escalation behavior, and workflow execution. Analytics enable organizations to:

• Measure AI effectiveness
• Monitor operational performance
• Identify optimization opportunities
• Validate automation ROI
• Ensure quality and governance

All analytics are Tenant-scoped and role-restricted.

---

​

Analytics Model Overview

Zild Assist tracks metrics across five domains:

1. Conversation Metrics
2. Agent Performance
3. Escalation & Human Intervention
4. Workflow Execution
5. Channel Performance

Each domain provides actionable insights for different stakeholders.

---

​

1. Conversation Metrics

Conversation metrics measure overall interaction volume and flow efficiency. Key metrics include:

• Total conversations
• Active conversations
• Closed conversations
• Conversation duration
• Messages per conversation
• Average response time

These metrics help evaluate:

• System load
• AI responsiveness
• Operational throughput

---

​

Conversation Lifecycle Tracking

Each conversation is tracked through:

• Created
• Active
• Escalated (optional)
• Closed

Lifecycle analysis helps identify:

• Long-running conversations
• High-friction flows
• Escalation bottlenecks

---

​

2. Agent Performance Metrics

Agent metrics evaluate AI effectiveness. Common indicators:

• Automated resolution rate
• Escalation rate
• Average response latency
• Tool execution frequency
• Tool success vs failure rate

High automation rates indicate effective knowledge and instruction configuration. Escalation spikes may indicate:

• Instruction ambiguity
• Knowledge gaps
• Workflow misconfiguration

---

​

Intent-Level Analysis

If intent classification is enabled, analytics can show:

• Most frequent intents
• Escalation by intent
• Conversion by intent
• Error-prone intents

This supports targeted optimization.

---

​

3. Escalation & Human Performance

Escalation analytics measure hybrid AI + human performance. Key metrics:

• Escalation rate
• Escalation reason breakdown
• Time to human response
• Conversation reassignment frequency
• Group-level workload distribution

Role visibility:

• Administrator → Full tenant visibility
• Supervisor → Group-level visibility
• User → Assigned conversations only

Escalation analysis helps balance automation and human intervention.

---

​

4. Workflow Execution Metrics

Workflow analytics measure operational automation. Tracked data includes:

• Workflow trigger count
• Workflow completion rate
• Failure rate
• Retry occurrences
• Execution latency

Examples:

• Leads created
• Contracts sent
• Tickets opened
• CRM updates performed

Workflow monitoring ensures deterministic processes operate reliably.

---

​

5. Channel Performance

Channel-specific analytics include:

• Conversations by channel
• Response time by channel
• Escalation rate by channel
• Conversion rate (for sales flows)
• Channel-level volume trends

This helps determine:

• Channel effectiveness
• Traffic distribution
• Operational pressure points

---

​

Sales & Conversion Analytics (AI SDR Use Case)

For sales-focused deployments, key metrics may include:

• Lead qualification rate
• Information capture completion rate
• Qualified lead volume
• Human handoff rate
• Conversion to meeting scheduled
• Conversion to opportunity

These metrics allow revenue teams to quantify AI impact.

---

​

Operational Dashboards

Supervisors and Administrators can monitor:

• Live conversation volume
• Active escalations
• Response latency
• Group workload
• Integration health

Dashboards enable real-time operational control.

---

​

Logs & Audit Visibility

Beyond aggregated metrics, Zild Assist provides:

• Conversation-level logs
• Tool execution logs
• Workflow execution logs
• Escalation history
• Configuration change tracking

This supports:

• Debugging
• Compliance
• Quality assurance
• Performance tuning

---

​

Identifying Optimization Opportunities

Analytics help identify:

• High-escalation intents
• Frequently failing tools
• Long conversation durations
• Underperforming workflows
• Channel bottlenecks

Optimization actions may include:

• Refining system instructions
• Improving knowledge sources
• Adjusting escalation rules
• Enhancing tool validation
• Redistributing group workload

Continuous improvement increases automation ROI.

---

​

Governance & Compliance Monitoring

Zild Assist analytics support governance by enabling:

• Audit trail review
• Permission change tracking
• Escalation oversight
• Configuration change monitoring

This ensures controlled AI deployment in enterprise environments.

---

​

Data Scope & Access Control

Analytics access follows RBAC rules:

• Administrator → Full analytics access
• Supervisor → Group analytics only
• User → Limited operational visibility

All analytics remain within Tenant isolation boundaries.

---

​

Best Practices for Analytics Usage

For mature deployments:

• Monitor automation rate weekly
• Track escalation trends
• Review workflow failure logs
• Compare channel performance monthly
• Conduct quarterly optimization reviews

Data-driven refinement improves long-term system performance.

---

​

Summary

Zild Assist analytics provide visibility across:

• Conversations
• Agent effectiveness
• Escalation behavior
• Workflow automation
• Channel performance

This allows organizations to:

• Measure AI impact
• Optimize automation
• Improve human collaboration
• Maintain governance
• Scale confidently

Analytics transform Zild Assist from a conversational tool into a measurable operational system.