Version 3.1
- 27 Mar 2024
- 1 Minute to read
- Contributors
- Print
- PDF
Version 3.1
- Updated on 27 Mar 2024
- 1 Minute to read
- Contributors
- Print
- PDF
Article summary
Did you find this summary helpful?
Thank you for your feedback
Breaking changes
Knovvu Analytics
- Analysis Worker service is now handling the text correction. Text normalization URL should be imported to Knovvu ML and its endpoint should be defined within the inventory.
- Knovvu ML should be upgraded to version 1.56 or higher. text_normalization service must be v9 and text_clustering v7 or higher.
- POST Tag method: tagId is added besides flagId. The method returns both tagId and flagId values. The flagId will be deprecated in 24 November 2023. Please use tagId instead.
Important Note: Required action after the update
- Click reorginise minIO files button under administration dashboard page.
- Remove the RabbitMQ analysis queue and recreate.
- Action required for AWS Updates : ML URL is required to be changed to ml.cloud.knovvu.com for AI Categorization and retrain for enhanced categories.
Enhancements
• Mixed language models: Mixed language models have been enabled.
• Creating waveform for meeting conversations.
• New query items: Customer sentiment, sentiment trend, agreement status
• Queries and topics have been unified: Category and topic descriptions have been merged. AI Topic management page have been moved under categories.
• New query item: Not interval query
• New API method: Patch user
• New evaluation criteria: Interaction count.
• Sentiment scores have been added to statistical comparison and analysis pages.
• Selection of Score type has been removed.
• Batch tagging
Improvements
• Security improvements
• Limitation updates for SR licensing: SR limit per agent per day is 12 hours.
• Email subject updates.
• Improvement in API performance
• UI language corrections
• Period selection is enabled for sentiment and NPS graphs.
• Performance improvements in excel export.
• The problem of duplication of forms has been resolved.
• The problem of duplication of categories has been resolved.
• The problem in seeding data while creating new tenants has been resolved.
• The problem in appearance of indexed categories has been resolved.
• The problem in saving forms with acoustical parameter-based question has been resolved.
Known Issues
- Bug: Excel problem in meeting conversations
- Bug: Customer sentiment in chats
- Bug: Precision normalization error
- Bug: Not showing the profile picture in profile
Was this article helpful?