Overview
The hTech AI Dashboard is the central control panel for the AI Ticket Assistant. It provides a full overview of AI activity across your support system, including insights, suggestions, routing behavior, analytics, logs, and queued replies. This guide walks through each dashboard section and explains how to use it effectively.
Main Sections of the Dashboard
The dashboard typically includes the following panels:
- AI Insights – Sentiment, complexity, escalation flags, and message summaries.
- Knowledge Suggestions – Potential KB improvements generated by the AI.
- Recent Logs – A timeline of AI actions, attempts, and safety blocks.
- Call Stats Overview – Token usage, model performance, and estimated cost.
- Queue Status – Delayed replies waiting to send.
- Top Ticket Tags – Common issues identified across your support environment.
Each section provides visibility into how the assistant is working behind the scenes.
AI Insights Panel
This panel summarizes recent ticket analysis activity. Insights may include:
- Sentiment (positive, neutral, negative)
- Complexity (simple, normal, complex)
- Escalation Recommended status
- Priority Boost if the message seems urgent
- Notes for Humans such as potential concerns
Use this section to understand which tickets may need attention from senior staff.
Knowledge Suggestions Panel
The AI may recommend new Knowledgebase articles or improvements based on recurring customer issues. This panel displays:
- Suggestion Type (new article, improvement, missing steps, etc.)
- Suggestion Text – The content the AI recommends
- Confidence Score – How strongly the AI believes the suggestion is useful
- Status – Pending, approved, or dismissed
You can manually delete suggestions or purge them entirely from this section.
AI Logs Panel
The Logs panel shows recent AI actions. Common entries include:
- "AI reply generated"
- "Auto-reply blocked due to escalation"
- "Ticket routed to Billing"
- "Tag assigned: login"
- "Queue entry created"
- "API error during reply attempt"
Each entry includes the timestamp and ticket ID. Logs are helpful for diagnosing unexpected behavior. You may delete individual logs or purge all logs.
Call Stats Overview Panel
This section summarizes AI usage:
- Total Calls for the month
- Model Usage Summary (Hosted, BYO, or both)
- Token Consumption including input and output tokens
- Runtime Performance averages
- Estimated Costs for BYO users
This helps you monitor resource consumption and adjust settings to control cost or performance.
Queue Status (AI Delay)
If delayed replies are enabled, AI messages are placed into the queue before they are posted. This panel shows:
- Ticket ID awaiting response
- Scheduled send time
- Status (queued, locked, error)
- Attempt Count for retries
You can monitor delayed messages to ensure the cron is running properly.
Top Ticket Tags
The dashboard displays commonly detected tags from recent tickets. This helps identify:
- Trending issues (e.g., “login”, “setup”, “billing”)
- Opportunities for new guides or SOP improvements
- Routing optimizations
Tags are automatically assigned by the AI when enabled.
Using Dashboard Actions
The dashboard may include administrative actions such as:
- Purge Logs – Clears all AI log entries
- Purge Suggestions – Removes all knowledge suggestions
- Delete Individual Rows – Useful for maintenance
These actions do not affect tickets themselves and are safe for routine cleanup.
Best Practices
- Check the Insights panel daily for escalated or complex tickets.
- Review Knowledge Suggestions weekly to improve your KB.
- Monitor Logs to ensure AI behavior matches expectations.
- Use Call Stats to track and control model usage and cost.
- Verify Queue Status if delayed replies are enabled.
- Use Top Tags to guide content creation and SOP updates.
Summary
The hTech AI Dashboard gives you full visibility into how the AI operates across your ticketing system. By reviewing insights, logs, suggestions, analytics, and queue activity, administrators can fine-tune automation, improve support efficiency, and maintain full control of AI behavior.
If you need help interpreting dashboard data or configuring optimal settings, feel free to open a support ticket.