4.96¶
Upcoming Feature Deprecations and Removals
- AudioCodes Nodes and Generic Voice Nodes have been deprecated. Flows with these Nodes continue to work, and you can still edit and clone them as well as export them as a Package. However, you can't create new ones in Cognigy.AI v4.96 or later.
- On February 15, 2024, Google deprecated Gemini 1.0 Pro. The shutdown date for this model is April 9, 2025. Google recommends migrating to Gemini 1.5 Pro and Gemini 1.5 Flash before April 9, 2025. We strongly recommend carrying out this migration since Gemini 1.0 Pro support will be removed from future Cognigy.AI versions.
- On April 29, 2025, we will discard
say
verbs withempty text content
affecting the following Voice Gateway Nodes:
Refer to the list of all upcoming removals on the Deprecations and Removals page.
Give Docs Feedback
Help us improve our product documentation on docs.cognigy.com
by sharing your thoughts in a quick survey. Your feedback shapes the future of our content!
4.96.3¶
Cognigy.AI¶
Released March 19, 2025
Bug Fixes¶
- Fixed the issue where Snapshots including Default Replies didn't work properly
4.96.2¶
Cognigy.AI¶
Released March 10, 2025
Bug Fixes¶
- Fix the issue where empty transcripts from the STT (speech-to-text) vendor caused an error in the Voice Copilot Endpoint
4.96.1¶
Cognigy.AI¶
Released March 5, 2025
Bug Fixes¶
- Fixed the issue where updating a Node using the API failed when trying to update fields with the same value
4.96.0¶
Cognigy.AI¶
Released March 5, 2025
Features¶
New AI Agent Voice Configuration¶
- Introduced a Voice Configuration section to both the AI Agent Creation wizard and the AI Agent Node, allowing users to fully customize voice settings, such as tone and language, for a specific AI Agent and its role. This customization ensures that the AI Agent's voice matches the intended personality, communication style, and tone appropriate for different use cases, enhancing the user experience and conversation quality while reflecting your product brand
Improvements¶
- Added a limit of 100 Cognigy.AI API calls from the Code Node to prevent excessive API usage
- Added caption messages when sending attachments with Genesys Cloud Open Messaging
- Added Claude 3.7 Sonnet to the list of available LLMs from Anthropic
Bug Fixes¶
- Fixed the issue where the Interaction Panel didn't show payloads including only data from untrusted Extensions
- Fixed the issue where updating an If Node resulted in a schema validation error
- Fixed the issue where the Knowledge AI chart on the Admin Center's Dashboard had the incorrect wording for Knowledge AI Chunks. The phrase
avaliable Knowledge AI Chunks
was replaced byconsumed Knowledge AI Chunks
- Fixed the Intent senctence generation process to ensure no extra information is included in the sentences
- Fixed the issue where a packaged Knowledge Chunk contained an embedding vector with reduced dimensions and resulted in a Package import failure
- Fixed the issue where the Adaptive Card generation failed for
gpt-4o
,gpt-4o-mini
, andgpt-35-turbo
- Fixed the issue where Nodes that use the Replace Special Phrases option for the Text Cleaner failed when replacing special characters
- Fixed the issue where the Add Memory Node didn't add a memory of a falsy boolean or numbers
- Fixed the issue where a new session created multiple Contact Profiles for the same contact
Cognigy Voice Gateway¶
Released March 5, 2025
Features¶
- Added the Duration Limit setting to the Transfer Node and Voice Gateway Endpoint when using dial. With this setting, you set the maximum duration of call, after which the call ends even if the caller is on the line
Improvements¶
- Improved the tooltip text of the Failover Transcribe Enabled setting in the Call Failover section of the Voice Gateway Endpoint settings
- Added support for Elevenlabs's model Flash v2.5 for the Synthesizer (Text-to-Speech)
- Added the capability to transfer the call when an error occurs with the TTS vendor of the Synthesizer (Text-to-Speech) or Recognizer (Speech-to-Text)
- Added support for Elevenlabs's model Flash v2.5 to
api-server
- Added the Voice over Internet Protocol (VoIP) carrier to the
X-Requested-Carrier-Sid
header for thesip:refer
verb in outbound calls - Added a maximum time-to-live (TTL) for sessions to ensure stale sessions are cleaned automatically. This enhances the overall system robustness and reduces potential memory leaks
Bug Fixes¶
- Fixed the issue where the model from the Synthesizer (Text-to-Speech) wasn't applied when using Elevenlabs as a TTS vendor
- Fixed the issue where a call's duration was set to
undefined
, which resulted in the errorValue is not a valid number: undefined
. An explicit validation forduration
was added before invokingmetricCallSessionDuration.set
when a call is completed - Fixed the issue where the Refer message could include a Session Initiation Protocol (SIP) contact address when the far end of a call used a Secure Inter-Session Protocol (SIPS). A requirement was added so that the Refer message includes a SIPS contact address rather than SIP
- Fixed the issue in
sbc-outbound
where outbound calls couldn't be transferred usingREFER
- Fixed the issue in sbc-outbound where interoperability occurred between SIP and SIPS (Secure SIP) communications. As a result,
REFER
messages now include a SIPS contact header when the far-end endpoint uses SIPS - Fixed the issue where the second
INVITE
request was sent to the incorrect socket during outbound calls - Fixed the issue where the occurrence of a
CALL_COMPLETED
event during a dial transfer turned all other events intoCALL_COMPLETED
orCALL_CREATED
- Fixed the issue where the Silence Overlay settings couldn't be saved in the Set Session Config Node
Cognigy Webchat¶
Released March 4, 2025
Improvements¶
- Added the Persistent menu to Webchat v3 and the corresponding settings to the Webchat v3 Endpoint settings
- Added the capability to handle typing indicators for Genesys Cloud Open Messaging and display when a human agent is typing in Webchat
- Added support for new Webchat v3 settings mode: Webchat v3 can be configured to render Markdown, collate streamed output messages, and display messages as progressive stream
Bug Fixes¶
- Fixed the issue where the message collation didn't show the header for empty messages
Cognigy Live Agent¶
Released March 5, 2025
Bug Fixes¶
- Fixed the issue where the email transcripts and notifications sent with the Send conversation by email function weren't displaying the
Powered by Cognigy.AI
label correctly due to missing data for brand name and incorrect logo
Cognigy Insights¶
Released March 5, 2025
Improvements¶
- Added the Content Security Policy header to the HTML of the Insights to prevent XSS (Cross-Site Scripting) attacks
- Renamed Goals Completed to Tasks Completed in the Contact Profile details and the Transcript Explorer for consistency and better alignment with the Task feature
- Added Endpoint icons to the Transcript Explorer
- Added new geographical analytics queries to track user location by city and country
- Added a tooltip with Channel information to the Endpoint icon in the Transcript Explorer and removed Channel column
- Added the
Top
andBottom
filter to the detailed view of the Goals performance chart, increased the displayed Goals from 5 to 10 in the Goals charts, added labels and a white border to the chart bars, and applied Goals filter to the Goals performance chart
Bug Fixes¶
- Fixed the issue where Goals metrics with a value of 0 were displayed on the legend of the Time Saved chart
- Fixed the issue where the Transcript Explorer showed an incorrect timestamp for time zones observing the daylight saving time
- Fixed the issue where the timestamp filters with time zone offsets weren't handled correctly in OData queries
- Fixed the issue where the Playbook runs failed with a
type
validation error - Fixed the issue where the Goals performance chart displayed the incorrect total count in the tooltip
- Fixed the issue where the Transcript Explorer detailed view displayed an error for time zones observing the daylight saving time
- Fixed the issue where the OData Steps Endpoint returned incomplete data. Added the missing
id
field to the OData Step data - Fixed the issue where Transcript Explorer broke if a Slot wasn't matched correctly
Infrastructure¶
Cognigy Voice Gateway¶
- Set FreeSWITCH
{{ .Values.freeswitch.args }}
to- --codec-answer-generous
:
FreeSWITCH prioritizes the incoming call's preferred codec, even if it's not the most efficient for FreeSWITCH. FreeSWITCH supports the following audio codecs:- PCMU (G.711 μ-law) - 64 kbps, low CPU usage, no compression
- PCMA (G.711 A-law) - 64 kbps, low CPU usage, no compression
- OPUS - Variable bitrate (6-510 kbps), excellent quality, higher CPU usage
- G722 - 64 kbps wideband, good quality, moderate CPU usage
Version Compatibility Matrix¶
The version compatibility matrix was updated for the following Cognigy products: