Uniphore Help Center Portal

In order to use the generative AI-based Call Summary feature, a Business Analyst must enable it for the associated AI Agent Assist  client application:

After enabling the GenAI Summary, a Business Analyst can configure how the Summary will be generated at runtime and how it will be formatted for each Experience.

After completing the configuration activity, the GenAI Summary is ready for use in the Agent AI Assist application. For more information about how the Agent uses the GenAI Summary during a call, click here.

The GenAI Summary can include the value of any Entity that has been mapped to the current Business Process. The Summary can be configured to display both detected Entities that were uttered by the Customer or the Agent during the call, as well as undetected Entities.

The Entities section of the GenAI Summary page lists all of the Entities that you want Real-Time Guidance Agent to detect during a call, and tells the AI Agent Assist application where and how they should appear in the Summary when it is displayed to an Agent.

After an Entity is added to this list, it can be edited or reordered as needed, as well as deleted.

At the top of the Entities list, click Add Entity. A new list row will appear.

Each listed Entity uses the following settings:

An Entity Label can be displayed before an Entity to clarify and help differentiate between Entities in the GenAI Summary.

When the Show Label switch is toggled On (green), the design preview, the Summary shown to the Agent, and the Conversation Summary variable that is saved with Summary submission will include the label(s) you have assigned:

To change the order of the listed Entities at any time, locate the Entity that you wish to move, click on the Reorder icon at the right side of the list, and drag the Entity to its desired position.

Changing the order of an Entity in the list also changes its order in the displayed Entities in the design preview, the Summary shown to the Agent, and the Conversation Summary variable that is saved with Summary submission.

Reordering can also impact which Entities are included in the minimum number of Entities that are shown to the Agent. By default, only the first four Entities will be displayed to a user before they click the Show more button.

Right-click on the Ellipse icon at the right side of the listed Entity and click Edit to open the Entity and change any of its settings. When you are finished, right-click again on the Ellipse icon and click Save.

Right-click on the Ellipse icon at the right side of the listed Entity and click Delete. The current Entity will be removed immediately from the Entities list; however, you must click Save at the lower right corner of the GenAI Summary page in order to finalize the deletion.

At runtime, GenAI Call Summary output can be produced in a target language that is different from the default language that is configured for an Experience.

Setting a different target language forces the Call Summary to be generated and submitted in the selected language for its call session output, regardless of which language is used in real time by the Agent and end user. For example, a CRM used by your account may require the target language for processing.

To change the target output language, at the bottom of the GenAI Summary configuration page, select the desired language from the Target Language dropdown.

In order to be available for selection in the dropdown, the desired target language must already be configured with the Business Process being used by the current Experience.

At run time, the GenAI Call Summary identifies short calls before they are sent to the LLM service and presents a notification message to an Agent indicating that no summary is generated due to the short duration.

Short-duration calls can be configured to identify short calls at run time and display a predefined custom message in the GenAI Call Summary when either the call duration is less than the specified duration or the number of speaker turns in the call is less than the specified turns.

The short-duration call uses the following settings:

Real-Time Guidance Agent leverages customized, industry-specific Summary Templates created in the Administration Platform to automatically generate a valuable and actionable AI-powered Call Summary to an Agent at the end of each call. This enables the Agent to view the contextualized call summaries that go beyond generic insights.

These Summary Templates, powered by Large Language Models (LLMs), support multiple languages and can be tailored to different industries. Multiple versions of industry-specific Summary Templates can be maintained for an account in the Administration Platform, based on the customer's requirements. For more information about defining the Summary Template in the Administration Platform, click here.

An Account Admin or Business Analyst can map a specific version of a Summary Template from the Administration Platform to an Experience. This mapped Template is considered as the default and automatically assigned at the beginning of a call. They can also assign specific versions of Summary Templates to Intents, allowing the system to dynamically apply the relevant template each time a new Intent is detected as the call progresses. For more information on viewing a customized GenAI Call Summary, click here.

A Flow Designer or Developer can optionally use the Summary Builder API to change the Summary Template while a call is still active, based on logic that has been defined in the Flow during the design phase. For example, the Flow design can call this API to switch to a different Summary Template whenever a specific Alert notification is received during the call. For more information on using the Summary Builder API, click here.

Real-Time Guidance Agent offers an advanced generative AI-based Call Disposition capability to automatically categorize calls using LLM zero-shot classification. Disposition categorization values can be set for an account in the Administration Platform based on the customer's requirements. Typically, this is a four-level category that describes the resolution of the call. For complete information about defining the Disposition Configuration in the Platform, click here.

The Call Disposition enables Real-Time Guidance Agent to present an automated and structured Disposition categorization to an Agent using a predefined Disposition Hierarchy in the Platform based on the call reason, call outcome, and/or customer's requirements.

A key benefit of the Real-Time Guidance Agent LLM-based Call Disposition is that it potentially eliminates the need for manual call categorization, saving an Agent's After Call Work (ACW) time and preventing call classification inconsistencies. It also helps to indicate next steps and provide information for processing and analysis downstream, improving insights into how the call reason was handled.

Disposition display is enabled on the Client Settings page for the AI Agent Application. The Agent can manually change the auto-generated Call Disposition values in the GenAI Call Summary, depending on the Call Disposition settings under the Client Settings page. For more details about setting Call Disposition on the Client Settings page Client Settings, click here.

A Business Analyst can map the relevant Disposition to an Experience, based on business needs. The Disposition categorization list appears in the GenAI Call Summary component of the AI Agent Assist application during runtime.

An Account Admin or Analyst can recover a failed Call Summary submission through a retry without data loss. To enable this, the system variable 'SummarySubmitServiceFlowStatus' must be used in both the Service Flow and Data Flow, and assigned to the Call Summary and Retry actions in the specific AI Agent Assist client application.

For more information on defining Service Flows to identify Summary submission failures, click here.

For more information on defining Data Flows for returning Summary submission retry results, click here.