2.9.0

Introducing a time limit for a call to remain in pause state, by default 10 minutes. This limit is used to overcome an issue with calls left on hold for very long periods.

Contact center flow now supports deviceprint enrollment and verification.

• Deviceprint enrollment triggers in the background after opting in for voiceprint enrollment or after voiceprint verification match.
• Deviceprint verification triggers when the voiceprint verify progress reaches the configured checkpoint for additional verification factors.

By default, the deviceprint factor is disabled.

Contact center flow triggers re-authentication when:

• Last session decision was a match
• Configurable duration passed from last session decision

By default, the re-authentication option is disabled.

This approach could help in scenarios where the person started the call and was authenticated but after some time a fraudster got control on the same call. The re-authentication results in restarting the audio and restarting the flow for the last identified person. The same Gatekeeper session is used to log the re-authentication related factors.

New scope-level configuration parameters for deviceprint are added to the Contact Center tab:

• DeviceprintTagKey
• DeviceprintTagNotFound
• EnableDeviceprint
• EnableDeviceprintDetectKnownFraudsters
• EnableDeviceprintEnrollment
• EnableDeviceprintKey
• EnableDeviceprintValue
• ReAuthenticateAfterMatchDelay

A support of spoof detection has been added to the post call processing.

We are introducing Caller ID Watchlist as a beta feature. This feature allows you to evaluate the risk level of a phone number based on a watchlist of known risky caller IDs and regular expressions rules. You can access the watchlist page to add, update, and delete your watchlist entries. You can also import multiple entries from CSV files. The feature is not fully integrated with the Web Agent Console yet, so you cannot use it to assess caller ID risk in real time. The public APIs are available in beta mode with limited support and documentation. The feature is still under development and testing, so you may encounter some issues.

The Caller ID Watchlist feature adds new operational configuration parameters under the CallerIdRiskAssessment tab:

• CallerIdRiskAssessmentFraudThreshold
• CallerIdRiskAssessmentReliabilityThreshold

The following new attributes have been added to the engagement’s exported CSV file:

• CallType
• RestartCounter
• Eligible Voiceprint Profile Enrollment
• Max Audio Reached for Voiceprint Profile Enrollment
• Min Audio Reached for Voiceprint Profile Enrollment
• Abandoned

A new filtering option on the Engagement page has been added based on a session attribute — Conveyed Session Decision to Agent.

The following new filtering attributes have been added to the engagement page:

• CallType
• RestartCounter
• Eligible Voiceprint Profile Enrollment
• Max Audio Reached for Voiceprint Profile Enrollment
• Min Audio Reached for Voiceprint Profile Enrollment
• Abandoned

Introducing a new online diagnostic tool that checks the connectivity between agent browser and Gatekeeper. This tool includes some tests related to browser settings, agent authentication (including token refresh), and GRPC streaming.

You can access this diagnostic tool at: https://[hostname]/diagnostic-tool/

In the Web Agent Console, a progress bar displays when detecting known fraudsters based on the current net audio divided by required audio for last checkpoint. It used to be based on checkpoints; as a result when having single checkpoint, the progress bar jumped from 0% to 100%.

It’s recommended to reduce the number of detect fraudster checkpoints back to one, if they were increased just to show some progress of the progress bar.

In the previous version, a webhook could only disable agent actions. Now, a webhook can enable agent actions such as opt in (enroll), opt out, and more.

The protocol supports a new action ENABLE_AGENT_ACTIONS with agentActionsToEnable parameter (list of actions to enable, same as for disabling agent actions).

Issue Description	Issue ID
Clicking a deep-linked URL to the Engagement doesn’t automatically select the Scope	XSUPPORT-8165
Inefficient execution time of the Fraud Time Machine	XSUPPORT-11539, XBIOMETRICS-33553
Errors when creating Kibana dashboards	XSUPPORT-12489
The scenario where engagement has no sessions isn’t handled correctly when saved in ElasticSearch	XSUPPORT-12655, XBIOMETRICS-33232
The voiceprint’s lock status isn’t always correctly displayed	XSUPPORT-12711
Clicking anywhere on the screen while uploading the audio or background model aborted the operation unexpectedly	XSUPPORT-12806
Cassandra Secure Sockets Layer (SSL) handshake exception	XSUPPORT-12956
The ListScopes API method isn’t working when using the REST endpoint	XBIOMETRICS-31314
Contact center sessions failed to start recording due to lack of recording ID. In certain cases, the lack of stop recording after pausing a recording caused timeouts when streaming the audio for post call operations.	XBIOMETRICS-32877

Enabling spoofing detection on orphan calls is considered beta.

Avoid the scientific notation, such as 1.15785E7, for long numbers in exported CSV.

Rarely it may happen that a post call session is isolated in a separate engagement and not aggregated with the other live sessions.

Gatekeeper supports NVSL 10 engine but:

• NVSL 10 can’t be used with the factory framework of the Risk engine 1.4 version.
• NVSL 10 frameworks are only used as baseline for the creation of new custom calibration model.

Deviceprint Public APIs are considered beta.

DEMO and Bootstrap risk engine calibration models might not work properly and they show poor accuracy. Use these models only for non-production use.

Neustar Factor integrated to the risk engine is considered beta.

Convoprint L2 Streaming functions for Business Logic Layer (BLL), Call Session Manager (CSM), and Web Agent Console (WAC) are considered beta.

Gatekeeper Web administrator applications don’t officially support Private Browsing Mode. Without them, the Console may display an authorization error. This isn’t required for agents authorized for a single scope.

The Session Person data (session grid) displays the person from the verify or identify voiceprint (not from convoprint), only if there is a match in those factors.

Possible misalignment between the text segment and the convoprint quality check.

The autocomplete for first name and last name doesn’t work when using special character, such as ‘@’ or ‘_’.

Web Admin Portal has a time zone selection issue related to browser updates.

Kibana Dashboard – The Generate PDF feature doesn’t work when Optimize for Printing is enabled.

NVSL TI_9 and TD_9 frameworks aren’t supported by the following bootstrap NRSL models:

• FACTORY_TI_RISK_ENGINE_BOOTSTRAP_1_2
• FACTORY_TD_RISK_ENGINE_BOOTSTRAP_1_2
• FACTORY_TI_RISK_ENGINE_DEMO_1_2
• FACTORY_TD_RISK_ENGINE_DEMO_1_2

A Clustering or Backwards Search action may time out due to the memory required for media collection exceeding configured limits. This is likely due to a collected engagement containing more than an expected number of sessions but could have another cause.

The following services require manual horizontal scaling:

• data-query-manager
• voiceprint-coordinator
• fraud-analytics-manager
• biometrics-supervisor
• transcription-coordinator

Clustering and Backward Search have the following limitations:

• No estimations available for media collection.
• Clustering and Backwards Search require manual scaling.

Audio streaming for transcription is limited to one realtime (this value isn’t configurable).

For Stereo audio, the redacted overlay interval applies on both channels in the audio viewer.

Realtime redaction isn’t supported when Transcribe API is used.