What's my plan?
Add-on AI agents - Advanced

Verified AI summary ◀▼

The Analytics dashboard, part of an early access program, offers a unified view of advanced AI agent performance. It consolidates existing dashboards into one, simplifying metrics analysis. You can filter reports by time frame, AI agent, channel, and more. The dashboard includes tabs for overview, contact reasons, and custom resolutions, helping you track key performance indicators and improve customer interactions.

The Analytics dashboard described in this article is in an early access program (EAP). If you aren’t part of the EAP, see About analytics for advanced AI agents instead.

In AI agents - Advanced, the Analytics dashboard provides a comprehensive, at-a-glance view of your organization’s overall advanced AI agent performance.

This article contains the following topics:

  • About the EAP
  • Accessing the dashboard
  • About the reports
  • Filtering the reports
  • Changing a report's view

About the EAP

Between August 8 and September 30, 2025, Zendesk is rolling out a refreshed analytics experience as an early access program (EAP). This EAP simplifies analytics for advanced AI agents so you can focus on the metrics that matter most.

The EAP is automatically rolling out to all customers over time. You’ll be notified via email approximately a month before your organization is added to the EAP.

When you’re added to the EAP, all of the existing analytics dashboards for advanced AI agents are consolidated into a new, single dashboard. This new dashboard is called the Analytics dashboard and is described below.

After you’ve been added to the EAP, it’s possible to temporarily revert to the previous analytics experience. To do so, contact your Zendesk account executive or contact Zendesk customer support. However, on September 30, 2025, the new analytics experience will become the default, and it will no longer be possible to access the previous experience.

Note: During the EAP, the Analytics dashboard is available in English only.

Accessing the dashboard

The Analytics dashboard is available under Analytics in the left sidebar.

To access the dashboard

  • In AI agents - Advanced, in the left sidebar, click Analytics.

About the reports

The Analytics dashboard contains the following tabs:

  • Overview tab
  • Contact reasons tab
  • Custom resolutions tab
Note: Conversations imported from third-party systems aren’t included in the Analytics dashboard.

Overview tab

The Overview tab shows key metrics about volume, overall AI agent performance, and BSAT scores. For the volume and performance metrics, results are also graphed over time and you can change the report’s view to group the results by AI agent, channel, or language.

You can filter the reports on this tab by AI agent, Channel, Reply type, Use case, Language, and Label.

Performance

The left side of the Performance section shows the following metrics:

  • Total conversations: The number of conversations handled by your AI agents.
  • Understood conversations: The percentage and number of conversations where an answer was provided from a knowledge source or conversations that were matched to a use case.
    Note: For AI agents with agentic AI, this excludes small talk and directly escalated conversations.
  • Escalated conversations: The percentage and number of conversations that were escalated from an AI agent to a human agent.
    Note: Escalation rate isn’t necessarily a direct inversion of Automated resolutions. One minus the escalation rate is the deflection rate, from which a subset of conversations are evaluated as potential automated resolutions.
  • Assisted conversations: Conversations where an AI agent took part but didn’t necessarily resolve the request. This includes:
    • Messaging conversations that weren’t escalated.
    • Emails where a use case was recognized, actions were taken, but no reply was sent.
  • Handled conversations: Conversations fully handled by an AI agent. This includes:
    • Messaging conversations with a recognized use case and no escalation attempts.
    • Emails where a reply was sent.
  • Automated resolutions: The percentage and number of automated resolutions. For details on how these are calculated, see Automated resolutions in advanced AI agents.
    Note: If you don’t have the Activate LLM setting enabled, this metric doesn’t appear.

The right side of the AI agent workload section shows the following BSAT metrics:

  • Rated conversations: The percentage and number of conversations where customers provided a response for BSAT, divided by the number of conversations where feedback was requested.
  • BSAT: The sum of the number of responses with a 4-5 score provided, divided by the total number of responses.
  • The colored bar indicates the proportion of each feedback response from 5 to 1, shown in shades of green, yellow, and red.
  • Below the bar, each rating is broken out with its respective percentage and number of responses relative to the total responses.

Custom resolutions tab

Note: This tab is available only to customers who deployed an AI agent on at least one messaging channel before September 8, 2025.

The Custom resolutions tab shows KPIs of each of the custom resolution states. For each of these KPIs, the results are also graphed over time and you can change the report’s view to group the results by total volume, channel, AI agent, or language.

You can filter the reports on this tab by AI agent, Channel, Reply type, Use case, Language, and Label.

Custom resolution state KPIs

The Custom resolutions tab shows the following custom resolution state metrics:

  • Total conversations: The number of conversations with an AI agent.
  • Custom resolution rate: The percentage and actual number of AI agent conversations that ended with a custom resolution state.
  • Informed: The percentage and actual number of AI agent conversations where the AI agent provided guidance to the customer. For more information, see Using the informed state.
  • Resolved: The percentage and actual number of AI agent conversations that ended with a meaningful resolution and no further questions from the customer.
  • Escalated via email: The percentage and actual number of AI agent conversations that were converted into an email ticket. This state is automatically set when the conversation flow reaches an Escalation block in a dialogue with the “Send an email” option set.
  • Escalated to agent: The percentage and actual number of AI agent conversations that were escalated to a human agent. This state is automatically set when the conversation flow reaches an Escalation block in a dialogue with the “Forward to an agent” or “Custom escalation” option set.
  • Undefined: The percentage and actual number of AI agent conversations that ended without a custom resolution state, often due to drop-off or use cases missing a custom resolution state.
  • Unresolved: The percentage and actual number of AI agent conversations where the issue was not answered or resolved.

Contact reasons tab

The Contact reasons tab shows key metrics about the performance of use cases and knowledge sources across your AI agents and channels.

From this tab, you can also jump to the use case suggestions report to quickly create new use cases that aren’t currently covered.

This tab contains the following reports:

  • Use cases
  • Knowledge sources

Use cases

The Use cases report shows a table of all existing use cases and their respective performance across your AI agents and channels. You can filter this report by AI agent, Channel, Reply type, Use case, Language, and Label.

This table includes the following columns:

  • Use case: The name of the dialogue or procedure.
  • AI agent: The AI agent that the dialogue or procedure was created for.
  • Channel: The channel that the AI agent is configured for. Possible values include Messaging and Email.
  • Reply method: How the reply was triggered. Possible values include Dialogues and Procedures.
  • Total conversations: The number of conversations where the use case was detected.
  • Automated resolutions: The number of automated resolutions across conversations where the use case was detected.
  • Custom resolutions: (Appears only for customers who deployed an AI agent on at least one messaging channel before September 8, 2025.) The percentage of conversations where a custom resolution was applied. For more information, see Custom resolutions tab.
  • Escalated conversations: The escalation rate across conversations where the use case was detected.
  • BSAT: The BSAT score across conversations where the use case was detected.

Knowledge sources

The Knowledge sources report shows a table of imported knowledge sources and their respective performance. You can filter this report by AI agent, Channel, Language, and Label.

This table includes the following columns:

  • Article name: The title of the article. This will be listed across all article sources included.
  • Knowledge base type: The knowledge source of the article. Possible values include Web crawler, Zendesk, and CSV.
  • Knowledge base name: The name the knowledge base was given when it was imported.
  • Usage: The percentage of conversations where the article was used to generate a reply (out of the total number of conversations where an AI-generated reply was provided).
  • Escalated conversations: The percentage of conversations where the article was used to generate a reply and the conversation was later escalated.
  • Automated resolutions: The percentage of conversations where the article was used to generate a reply and the conversation counted as an automated resolution.
  • BSAT: The average BSAT score across conversations where the article was used to generate a response.

Filtering the reports

You can filter the dashboard’s reports by time frame or other attributes.

To filter the dashboard by time frame

  1. In the Analytics dashboard, in the upper-right corner, click Timeframe.
  2. Using one of the following methods, select the dates you want to view conversations for:
    • On the right, select one of the predefined time frames: Today, Yesterday, Last 7 days, Last 30 days, This month, Last month
    • Select specific beginning and end dates on the calendar

To filter the dashboard by other attributes

  1. In the Analytics dashboard, click one of the following filters and select its values as needed depending on the data you want to see:
    • AI agent: The AI agent who participated in the conversation.
    • Channel: The channel the conversation took place in. Values can be Messaging or Email.
    • Reply type: The type of replies the AI agent can provide. Possible values include Use case, Knowledge, and Hybrid. (Not available in the Knowledge sources report.)
    • Use case: The use case or intent detected during the conversation. (Not available in the Knowledge sources report.)
    • Language: The language detected during the conversation.
    • Label: The labels applied to the conversation.
  2. Click Apply.
  3. Repeat the steps above to apply any additional filters as needed.

Changing a report's view

For the AI agent workload and custom resolutions reports, you can apply different views that change how the results are presented. Different views let you focus on different aspects of the overall data.

To change a report’s view

  1. In the Analytics dashboard, select the one of the following tabs:
    • For AI agent workload reports: Overview
    • For custom resolutions reports: Custom resolutions
  2. Select one of the metrics.
  3. Above the visualization on the left side, click the drop-down to select one of the following report views:
    • By channel
    • By AI agent
    • By language

  4. (Optional) Above the visualization on the right side, click Daily, Weekly, Monthly, or Quarterly to change how the report’s results are aggregated.
Powered by Zendesk