Work with an interaction reprocessing job

AI-generated summary

Genesys Cloud's interaction reprocessing functionality enables retroactive application of topic model changes to historical interactions, supporting organizations using CX 1/2 with WEM or STA Add On, CX 3, CX 4, or Cloud EX AI. The system processes up to 250,000 interactions per job, requiring specific SpeechAndTextAnalytics permissions and generating CSV outputs for newly detected topics. Jobs progress through Queued, In Progress, and Completed states, with comprehensive monitoring capabilities including progress tracking and filter details. Users access the feature through Admin or Menu interfaces, with options to view, export, and delete jobs. The platform includes a monthly quota of 2,000 interactions per agent, with additional charges for exceeding this limit. Best practices include focusing on programs with updated topics, applying specific filters, and maintaining records through interaction ID exports. While jobs cannot be edited after publication, they can be deleted and recreated if necessary. The system accommodates normal processing delays during Content Search data retrieval and provides detailed CSV exports for auditing and investigation purposes.

Feature coming soon. Billing is currently inactive during Limited Availability (LA); usage is free until GA release.

PrerequisitesClick to expand

Genesys Cloud CX 1 or Genesys Cloud CX 2 with WEM Add On or STA Add On, Genesys Cloud CX 3, Genesys Cloud CX 4, or Genesys Cloud EX
AI Experience tokens
SpeechAndTextAnalytics > ReprocessInteractions > Add permission
SpeechAndTextAnalytics > ReprocessInteractions > Delete permission
SpeechAndTextAnalytics > ReprocessInteractions > Edit permission
SpeechAndTextAnalytics > ReprocessInteractions > View permission
SpeechAndTextAnalytics > Programs > View permission
To view approximate count of interactions to be reprocessed, one of these permissions:
- Analytics > ConversationDetail > View
- Recording > Recording > View or Recording > RecordingSegment > View

This feature enables Genesys Cloud administrators and analysts to apply changes made to topics across historic interactions retrospectively . Use the reprocessing interface to create and manage jobs that retroactively analyze interactions based on your latest topic models.

Notes:

Scope of analysis

Reprocessing evaluates historical interactions based on the programs selected in the dropdown.
New programs with new or updated topics are not included, since historical interactions are not linked to them.
To include new topics, add or update them within an existing program.

Processing limits

Each reprocessing job can include up to 250,000 interactions.

Topic detection rules

If an interaction contains no detected topics, it is marked as failed during reprocessing.
If no transcript within an interaction contains detected topics (even if a transcript block does), the interaction is still marked as failed.
If the selected program has no topics at the time of reprocessing, the interaction’s existing topics remain unchanged.

CSV output behavior

The generated .csv file lists only newly detected topics.
It does not show removed topics or provide before-and-after comparisons.

View and understand the reprocessing interactions listClick to expand

Navigate to Admin > Interaction Reprocessing.
Click Menu > Conversation Intelligence > Speech and Text Analytics > Interaction Reprocessing.
A table lists all Reprocessing Jobs with statuses such as Queued, In Progress, and Completed.
For each job, you can:
- View details (filters, job creation time, program associations).
- Export impacted Interaction IDs.
- Delete completed or obsolete jobs.
Click the job name to view:
- Program selection
- Filter criteria
- Timeline range
- Media types and other attributes

Create a reprocessing interactions jobClick to expand

Navigate to Admin > Interaction Reprocessing.
Click Menu > Conversation Intelligence > Speech and Text Analytics > Interaction Reprocessing.
Click Create a Job.
Provide the following required inputs:
- Job name
- Date range (for the historic interactions to process)
- Programs (where Topic/Category changes apply)
- Media types (for example, voice, chat)
(Optional) Add advanced filters to refine your interaction selection.
Review the Estimated Interaction Count that appears.
Click Publish to submit the job for processing.

Note: Only topics linked to selected programs are reprocessed. Be sure to update or assign topics to relevant programs before starting reprocessing.

Monitor a reprocessing interactions jobClick to expand

After publishing, the job enters a Queued state.
Once started, a Progress Bar indicates the status:
- Queued – Awaiting a turn in the processing queue.
- In Progress – Processing interactions with real-time updates.
- Completed – Job is done; export option becomes available.
Hover over the progress bar and view details such as:
- Percentage complete
- Number of interactions processed
If the job is delayed, it’s likely due to the time needed to retrieve data from Content Search. This behavior is expected and typically resolves within minutes.

Export reprocessing interaction job resultsClick to expand

Understanding usage and billingClick to expand

Best practicesClick to expand

Practice	Description
Program awareness	Only include programs that contain updated topics.
Scope management	Apply tight filters (for example, media types, recent dates) to limit processing.
Export IDs	Download interaction IDs for records or further processing.

Common use case exampleClick to expand

FAQsClick to expand

Q: Can I edit a job after I publish it?
A: No. Once a job is published, it cannot be modified. You can delete and recreate it if necessary.

Q: Why can’t I export interaction IDs until a job completes?
A: Export is only available after processing is finished to ensure accurate results.

Q: How do I know which Programs contain my Topic?
A: Review Topic assignments in the STA configuration interface before running the job.

[NEXT] Was this article helpful?

Get user feedback about articles.

Work with an interaction reprocessing job View summary

[NEXT] Was this article helpful?

Work with an interaction reprocessing job