For the complete documentation index, see llms.txt. This page is also available as Markdown.

Insights

View analytics for your requests, tickets, and GPT-Powered Answers

Our Insights feature offers detailed reporting to track request load, response times, and service levels. Use the 'Request' tab for agent-specific insights and the 'Virtual Agent' tab for Clearfeed Virtual Agent data. Metrics can be filtered by channel, responders, collections, and more. They can also be viewed within specific working hours, with options to download granular data for deeper analysis.

Use Insights to answer questions like:

  • Which channels experience the highest request volumes?

  • How quickly are responders resolving requests on average?

  • How effectively is the Clearfeed Virtual Agent reducing manual workloads?

Insights page on the ClearFeed Web Dashboard

Supported Metrics

We support metrics across Requests, Tickets, and Answers. You can:

  • Select the metrics you want to track.

  • Break them down by parameters such as Collection, Channel, Assignee, Custom Fields, Customer, Customer Owner, Customer Custom Fields and more.

  • Apply a date range to analyze performance over time.

You can break down any metric using up to 3 dimensions, available in tabular reports only.

For All Requests/Tickets

Use the toggle Request with a ticket to see metrics for tickets (this is only applicable for users under Product Edition External Helpdesk and Integrations).

Number of Requests

  • Total requests initiated during the selected period; excludes earlier requests with messages in the period.

  • Toggle 'Requests with Tickets' to view insights for ticket-based requests.

  • Example: 10 requests initiated today, even if work continued on 4 older requests, the count is still 10.

  • Reporting Unit: Number

Number of Messages

  • Total Slack messages (from responders, requesters, or others) within the selected time frame.

  • Excludes bot messages not whitelisted by ClearFeed.

  • Includes messages from requests initiated before the selected period.

  • Reporting Unit: Number

Time to First Response

  • Average time from request creation to the first responder’s reply (e.g., 👀, 🎫).

  • Excludes open requests with no response; only counts responses within business hours.

  • Example: A 9:00 AM request responded at 9:30 AM has a 30-minute first response time.

  • Reporting Unit: Average in minutes.

Response Time

  • Average time between a requester’s message and the next responder reply, including first and subsequent responses.

  • Excludes open requests with no responses; business hours (support pending).

  • Example: If a first response is 4 mins and next is 8 mins, average response time = 6 mins.

  • Reporting Unit: Average in minutes.

Resolution Time

  • Average time from request initiation to marking it as "Solved," excluding unresolved requests.

  • Adjusted for business hours to exclude non-working periods

  • Adjusted to omit time spent in Pending, On Hold, or both states when Resolution Time pausing is configured in Business Schedule & SLA.

  • First Response Time is never paused by this setting; only Resolution Time is affected.

  • Example: A request raised at 8 AM and solved at 9:30 AM on a 9–5 schedule = 30 mins resolution time.

  • Reporting Unit: Average in minutes.

First Resolution Time

  • Time from request creation to its first resolution, unaffected by subsequent reopening.

  • Excludes unresolved requests; adjusted for business hours.

  • Example: A 10:00 AM request solved at 10:30 AM = 30 mins first resolution time.

  • Reporting Unit: Average in minutes.

Resolution Rate

  • The percentage of requests that have been marked as resolved out of the total requests received.

  • Formula: (Number of Closed/Resolved Requests ÷ Total Requests) × 100

  • Example: If 80 out of 100 requests were resolved, the Resolution Rate = 80%.

  • Reporting Unit: Percentage (%)

CSAT Metrics

Note: CSAT metrics are aggregated based on the time the survey is triggered (i.e., when the request is marked as solved) - not when the request was created or when the response was submitted.

CSAT 5-Point Survey

Available when CSAT Type is set to “Emoji Rating”

  • Displays the average satisfaction score (1–5) from requesters for solved tickets.

  • Requests without CSAT feedback are excluded.

  • Toggle ‘Requests with Tickets’ to view only ticket-based scores.

  • Example: Ratings of 1, 3, and 5 result in an average score of 3.

  • Reporting Unit: Average score

CSAT 2-Point Survey - Positive Responses

Available when CSAT Type is set to “Thumbs Up/Down Rating”

  • Shows the count of positive (👍) responses from requesters on solved tickets.

  • Toggle ‘Requests with Tickets’ to view only ticket-based scores.

  • Reporting Unit: Count of positive responses

CSAT 2-Point Survey - Negative Responses

Available when CSAT Type is set to “Thumbs Up/Down Rating”

  • Shows the count of negative (👎) responses from requesters on solved tickets.

  • Toggle ‘Requests with Tickets’ to view only ticket-based scores.

  • Reporting Unit: Count of negative responses

SLA Metrics

% of First Response SLA Breached

  • % of requests breaching the SLA for first response, adjusted for business hours.

  • Example: 10 requests with 6 breaches = 60% breach rate.

  • Reporting Unit: Percentage (%).

% of Resolution Time SLA Breached

  • % of requests breaching SLA for resolution time, excluding unresolved or non-request messages.

  • Example: 10 resolved requests with 6 breaches = 60% breach rate.

  • Reporting Unit: Percentage (%).

% One Touch Resolution Breached

  • % of requests breaching "One Touch SLA" (e.g., resolution in a set number of replies).

  • Excludes non-request messages; unaffected by business hours.

  • Example: 10 solved requests with 6 breaches = 60% breach rate.

  • Reporting Unit: Percentage (%).

Answers Metrics

Number of Interactions

  • Tracks how often the Virtual Agent is triggered, including responses (generated answers, search results) and no-response instances.

  • Each trigger counts as one interaction, even if it includes multiple response types.

  • Example: 10 triggers = 4 generated answers + 3 search results + 3 no responses; total = 10 interactions.

  • Reporting Unit: Number.

Number of Responses

  • Captures instances where the Virtual Agent responded with generated answers or search results.

  • Excludes triggers where no response is found.

  • Example: 10 triggers = 4 generated answers + 3 search results + 3 no responses; total = 7 responses.

  • Reporting Unit: Number.

Number of Generated Answers

  • Counts instances where the Virtual Agent responded specifically with generated answers.

  • Excludes responses from search results or no-response triggers.

  • Example: 10 triggers = 4 generated answers; total = 4 generated answers.

  • Reporting Unit: Number.

Number of Positive Feedback

  • Tracks how often users selected "Yes, marked as solved" for Virtual Agent solutions.

  • Excludes instances with no feedback provided.

  • Example: 10 triggers = 5 positive + 2 negative feedback + 3 no feedback; total = 5 positive feedback.

  • Reporting Unit: Number.

Number of Negative Feedback

  • Tracks how often users selected "No, create a ticket" for Virtual Agent solutions.

  • Excludes instances with no feedback provided.

  • Example: 10 triggers = 5 positive + 2 negative feedback + 3 no feedback; total = 2 negative feedback.

  • Reporting Unit: Number

Virtual Agent Metrics

ClearFeed exposes the following metrics to help you measure AI Agent performance, effectiveness, and customer satisfaction:

Number of Sessions

Total count of AI Agent sessions that were created. This represents the overall volume of AI interactions and is used for billing.

Number of Sessions where Virtual Agent Answered

Sessions where the AI Agent provided a response intended to resolve the user’s request.

Answer Rate

Percentage of AI Agent sessions where the AI Agent provided an answer intended to resolve the user’s request.

  • Formula: Answer Rate = (Number of Sessions where Virtual Agent Answered ÷ Number of Sessions) × 100

  • Example: 80 answered sessions out of 100 total sessions = 80% Answer Rate.

  • Reporting Unit: Percentage (%).

  • If there are no AI Agent sessions in the selected period, this metric is shown as N/A.

Number of Deflected Sessions

Sessions where the AI Agent successfully deflected the request. A session is marked as deflected when the user clicks “Yes, mark as solved.”

Deflection Rate

Percentage of AI Agent sessions that were successfully deflected.

  • Formula: Deflection Rate = (Number of Deflected Sessions ÷ Number of Sessions) × 100

  • Example: 60 deflected sessions out of 100 total sessions = 60% Deflection Rate.

  • Reporting Unit: Percentage (%).

  • If there are no AI Agent sessions in the selected period, this metric is shown as N/A.

Number of Sessions with Positive Feedback

Sessions where users provided positive feedback by selecting “Yes, mark as solved” or “Yes, helpful.”

Number of Sessions with Negative Feedback

Sessions where users selected “No, create a ticket” or “Not helpful”, indicating the AI response did not resolve their need.

Aggregation

  • You can aggregate metrics in three different ways for detailed reporting and insights:

    • Average

    • Median

    • Percentile: Metrics can be aggregated for the 75th, 90th, and 95th percentiles.

  • Aggregation options are available for the following metrics:

    • First Response Time

    • Response Time

    • First Resolution Time

    • CSAT Score

  • For other metrics, the default aggregation method is Average.

Supported Aggregation in ClearFeed

Breakdown By and Distribute By

Breakdown By—Split any metric by single-select fields only: Assignee, Status, Priority, Collection, Customer, single-select custom fields, or native AI fields (Auto-CSAT, Auto-Emotion). Not available for multi-select fields. Use up to 3 dimensions in tabular reports.

Distribute By—Distribute request counts across multi-select fields, Available only for the Number of Requests metric. Available fields: Contributors, Auto-Category, or multi-select custom fields. One request can appear under multiple values, so sums may exceed total requests.

Insights report with Distribute By set to Contributors
Insights report with Distribute By set to Contributors

When Distribute By is enabled, Breakdown By is disabled, and vice-versa.

Export Data

ClearFeed supports data export/download in CSV format. The data can be exported by clicking the button highlighted below:

To export the data via APIs, the following methods can be used:

Standard Reports

ClearFeed creates a set of standard Insights reports by default so teams can quickly open commonly used metrics without building reports from scratch. These reports appear under Shared Reports and use the last 7 days as the default date range.

On the Request tab, default reports can include:

  • Open Requests: Open request volume over time.

  • Requests by Assignee: Request volume grouped by assignee.

  • Requests Volume: Request volume grouped by customer for customer inboxes, or by channel for other helpdesks.

  • Avg First Response Time: Average first response time over time.

  • Avg Response Time: Average response time over time.

  • First Response SLA Breach: First response SLA breach percentage, shown when first response SLA tracking is configured.

  • CSAT Score: Customer satisfaction score or positive CSAT responses, shown when CSAT is enabled.

On the Virtual Agent tab, default reports can include:

  • Sessions: AI Agent session volume over time.

  • Deflection Rate: Percentage of AI Agent sessions that were deflected.

  • Negative Feedback Sessions: AI Agent sessions with negative feedback.

The exact standard reports available in your account depend on your ClearFeed edition and configuration. For example, SLA and CSAT reports appear only when those features are configured, and integration-focused accounts may see a generic Default report instead.

Custom Reports

You can save your frequently used filter combinations as a custom report to access specific data without reapplying complex filters each time. Start from a standard report, an existing custom report, or a new report setup.

Creating a New Custom Report

You can create a new custom report from scratch or use an existing report as the starting point.

Option 1: Using an existing Report

  1. Start with an existing standard or custom report and update it by:

    • Changing the metric

    • Changing the breakdown

    • Editing or removing existing filters

    • Changing the date range

  2. Once the desired filters are applied:

    • Click Save as New Report to create a new report

    • Click Save Report to update the existing report

  3. In the Save As overlay that appears:

    • Enter a name for your new report

    • Choose the visibility (Admins only):

      • My Reports: Visible only to you

      • Shared Reports: Visible to all users in your account

  4. Click Save to create the report

  5. You will be redirected to your newly created report

Option 2: Clone an Existing Report

  1. Identify the report you want to clone

  2. Click the three dots () next to that report

  3. Select Clone Report

  4. In the popup:

    • Enter a name for the new report

    • Choose the visibility

  5. The cloned report will be created, and you will be redirected to it

  6. Optionally, edit the filters or make changes to the report as needed and click Save Report

Managing Custom Reports

For Reports Created by You

If you have created a custom report, you can:

  • Rename: Change the name of the report.

  • Clone: Duplicate the report to create a new one.

  • Delete: Remove the report permanently.

For Shared Reports You Didn’t Create

For reports under Shared Reports that you didn’t create, you can:

  • Clone: Use the existing report as a template to create your custom report.

FAQs

  1. Does Insights support timezone customization for reports? Answer: Yes, all reports in Insights follow the default timezone set in Account Settings. You can update the timezone anytime by navigating to Account Settings → SLA Settings.

  2. Which type of custom fields are available for Breakdown By and Distribute By? Answer: Use Breakdown By with single-select custom fields. Use Distribute By with multi-select custom fields for the Number of Requests metric.

  3. Does Resolution Time calculation exclude time spent in Pending or On Hold? Answer: Yes, if Pause Resolution Time is enabled under Settings → Additional Settings. Admins can configure ClearFeed to exclude time spent in Pending, On Hold, or both states from Resolution Time. This pause setting does not affect First Response Time.

Last updated