AI logs in Thena The AI logs tab is a dedicated view inside each ticket, shown alongside the Conversation, Notes, and Activity tabs. It captures every action taken or attempted by AI agents, providing a transparent audit trail of automated activity. This tab ensures full visibility into how AI contributes to ticket handling—from small updates to high-impact decisions.

​
Why it matters

AI agents can streamline and accelerate customer support, but transparency is non-negotiable. The AI logs tab gives your team a reliable source of truth for everything the AI touches—making it easy to review, verify, and trust automation at scale. It’s not just about documenting actions; it’s about accountability and explainability.

​
What gets logged

AI logs are designed to record everything an AI agent does—from operational tasks to decision-making attempts. The current examples include:
  • Status changes
    When the AI updates the ticket status, such as moving it to Pending customer reply.
  • Fallbacks to human agents
    If the AI defers to a human due to a policy constraint, confidence issue, or unknown intent.
  • Custom field updates
    Setting fields like “Issue category” or product tags.
  • Ticket associations
    Linking related tickets to provide context or history.
  • Summaries and notes
    When the AI generates internal notes or conversation summaries.
These are just the beginning. As AI agents evolve, this tab will continue to capture a growing range of intelligent behaviors—such as automated escalations, workflow triggers, multi-language responses, proactive outreach, and more. If the AI does it, it gets logged here. Each entry includes who (or what) acted, when it happened, and often, why—via the Show reasoning link.

​
Show reasoning

When enabled, the Show reasoning option reveals the AI’s internal decision path—whether it was a confidence threshold check, a policy application, or a fallback rationale. This gives support managers and admins the insight they need to improve AI performance and tune behavior over time.

​
Use cases

Operational oversight

Track and audit AI performance just like you would with human agents.

Compliance and review

Meet auditability standards with a log of who did what, and why—even when “who” is an AI.

Debugging and tuning

Surface decision-making blind spots or overly strict policies that block automation.

Agent onboarding

Help new support reps quickly understand the AI’s activity before jumping into a ticket.