---
sourceDocument: Brazil Conversational Interfaces
sourceDocumentLink: https://www.servicenow.com/docs/r/conversational-interfaces

 Release :

    - brazil

ft:locale :

    - en-US

ft:publication_title :

    - Brazil Conversational Interfaces

ft:clusterId :

    - convint

bundleId :

    - convint

workflow :

    - Platform


---

# Insights into failed scenarios

# Insights into failed scenarios {#ariaid-title1}

Release version: Brazil  
Updated October 5, 2026  
![](https://www.servicenow.com/docs/portal-asset/ico-clock) 1 minute to read  
Analyze failed scenarios in an automated evaluation and see the recurring failure patterns behind them, to prioritize what to fix without reviewing every failed scenario one by one.

## View top failure patterns {#severity-insights__section_view_top_failure_patterns}

After an automated evaluation with failed scenarios finishes running, Assistant Designer analyzes the failed scenarios and clusters them into failure patterns. The Top issues detected section on the evaluation results page lists these patterns so you can see
where your assistant is breaking down across the whole data set.

By default, up to four failure patterns are shown, ranked by severity and then by frequency, from highest to lowest. You can:

* Sort the list by Severity or Frequency.
* Select Show issues to expand the list and view every failure pattern that was detected.
* Select View scenarios next to a failure pattern to see the specific scenarios where it occurred.
{#severity-insights__ul_top_issues_actions}

## View scenarios for a failure pattern {#severity-insights__section_view_scenarios_for_pattern}

Select View scenarios from either the Top issues detected list or a metric's issue chart to open the Scenarios list, filtered to the scenarios where that failure
pattern occurred. If you select more than one failure pattern, the list shows the scenarios where any of the selected patterns occurred.


The list includes:

* Scenario: The scenario name, which links to the scenario's transcript.
* Issues detected: The failure patterns found in that scenario, grouped by metric category.
{#severity-insights__ul_scenario_list_columns}

By default, scenarios are sorted by severity and then alphabetically. Select a column header to sort by Scenario or Issues detected instead.

## View a scenario transcript {#severity-insights__section_view_scenario_transcript}

Select a scenario name from the Scenario list to open its transcript.


The transcript page shows:

* The full conversation transcript. Select Download to save the transcript as a CSV file.
* Why was this flagged?: Every issue detected in the scenario, along with an explanation, sorted by severity from highest to lowest.
{#severity-insights__ul_transcript_page}  
Note:  
An issue tied to a per-conversation metric isn't associated with a specific turn in the transcript. An issue tied to a per-turn metric is labeled at the turn where it occurred, when possible.

Select Back to list of scenarios to return to the filtered scenario list.
**Related concepts**   

* [Testing assistant conversations](https://www.servicenow.com/docs/gE9qJJ_3Lo0fJ3LZSa88mw "The Testing tab in Assistant Designer helps you assess the quality and performance of your conversational assistant. You can test conversations manually or run automated evaluations on generated scenarios to measure the assistant's success in addressing the user request and usage of available assets.")
* [Customizing severity levels](https://www.servicenow.com/docs/WhBKzwCOu4yMkWjaysCYjQ "Assign a severity level to each failure pattern that automated evaluations can detect, so your evaluation results reflect how your organization wants to weigh each type of failure.")  
**Related reference**   

* [Evaluation metrics for automated evaluation](https://www.servicenow.com/docs/hWlKAN0kPz_w2TfN~Nkdag "Metrics used to define success for evaluations.")