Enlighten AutoSummary

Enlighten AutoSummary automatically generates a summary for agents at the end of inbound and outbound voice interactions. The generated summary appears in the Notes field in CXone Agent applications at the end of the interaction. It can be passed to a supported CRMClosed Third-party systems that manage such things as contacts, sales information, support details, and case histories. application, which makes the summary available for future agents who interact with the contact. The summary data can also be used in Interaction Analytics.

AutoSummary uses NICE Enlighten AI to analyze and summarize the interaction. The summaries capture key information about the interaction, such as: 

The summaries that AutoSummary generates are consistently concise, accurate, and easy to read and understand. Summaries added by agents may not always meet these standards. For example, an agent might add a summary that says: 

"CXJames Smith 345456/concerned about missing pymnt/DM researched/no record in acc/informed not posted/prms cb." 

The summary generated by AutoSummary for the same interaction might be:

"Mr. James Smith's last contact about Billing: Missing Payment was 7 days ago, had negative sentiment and was not resolved. Mr. Smith was offered and declined a bill credit. Was given a promised callback."

Agents can modify the generated summary if they need to add or correct details. This allows them to focus on the important parts of the interaction instead of trying to capture the entire interaction themselves.

AutoSummary is an agent assist application that's set up and managed in Agent Assist Hub.

How AutoSummary Works

AutoSummary uses NICE Enlighten AI. It also uses Generative AI, which is based on Large Language Models (LLM) technology. Generative AI models learn patterns and the structure of data during their training. They are then able to generate new data that follows similar patterns.

Diagram showing call audio being transcribed in TTS, then processed by the AutoSummary model and outputting summary text.

AutoSummary provides several models that are trained to recognize terminology, intentsClosed The meaning or purpose behind what a contact says/types; what the contact wants to communicate or accomplish, actions, and outcomes related to various vertical markets. These include markets such as financial services, healthcare, telecommunications, travel and hospitality, and retail. Combined with Generative AI and LLM technology, this allows AutoSummary to produce short, readable, and accurate summaries.

During inbound or outbound voice interactions, audio from the call is captured and transcribed to text. The custom Studio script that routes interactions is configured with an AutoSummary profile. The profile contains settings that determine the model to use and the kind of summary to generate.

AutoSummary applies the profile settings to the transcription. The model identifies the key aspects of the interaction. These include the contact's intentClosed The meaning or purpose behind what a contact says/types; what the contact wants to communicate or accomplish, sentimentClosed Overall mood or result of the interaction as determined by analysis of words, phrases, and context of the transcript., actionsClosed Events or actions that ocurred during an interaction, such as making a return or completing payment. that occurred during the interaction, and the interaction's outcomeClosed Result assigned by the agent or system at the end of a voice (disposition) or digital (status) interaction.. The result of the analysis is sent to the Generative AI engine for summarization.

AutoSummary delivers the summary to the agent in CXone Agent or one of the other supported agent applications. The resulting summary data can be delivered to a CRMClosed Third-party systems that manage such things as contacts, sales information, support details, and case histories. or used in Interaction Analytics.

Summary Types

You can choose the type of summary that AutoSummary produces for each AutoSummary profile. All summaries are AI-generated based on the interaction transcript. The types are:

  • Structured AI: This type of summary defines the intent, outcomes, and events from the interaction in short phrases. For example:

    Customer contacted about Making A Payment.

    The customer's sentiment was Positive and the issue was Resolved.

    Outcomes that occurred during the contact: Payment Completed

    Actions: Payment was Made, Balance Payment was Processed, Client Made Payment

  • Generative AI: This type of summary comes directly from the Generative AI LLM engine. It consists of excerpts of the interaction transcript that summarize the intent, outcomes, and events from the interaction. The following is an example of a Generative AI summary:

    The customer attempted to make a bill payment but encountered issues due to an incorrect routing number. The customer attempted to make the payment on time. The agent collected the correct routing number from the customer. The agent offered to investigate the payment. The customer requested assistance with making the payment, exhibiting a positive sentiment throughout the call. The payment was processed and the customer made the payment successfully.

  • Hybrid: To create this type of summary, AutoSummary sends sentences related to the Structured AI summary to the Generative AI service. The service provides a summary in complete sentences. The following is an example of a hybrid summary:

    The customer encountered issues with a bill payment due to an incorrect routing number. The agent confirmed the correct routing number and offered to investigate the payment. The customer requested assistance with making the payment, exhibiting a positive sentiment throughout the call.

AutoSummary Data in CXone

The interaction summaries that AutoSummary generates are passed to the agent application. To save the data, you must configure AutoSummary to pass it to your CRMClosed Third-party systems that manage such things as contacts, sales information, support details, and case histories.. This makes summaries from previous conversations with the contact available in the CRM. Other agents who handle interactions with the contact can quickly understand the contact's history with your organization. Without a CRM integration, the generated summaries aren't visible to other agents.

AutoSummary data is also available In Interaction Analytics. If you use that application, you can filter interactions based on the intents, actions, and outcomes captured in generated summaries.

Agent Experience with AutoSummary

Agents don't need to do anything to make AutoSummary work. After it's set up, AutoSummary automatically generates summaries and sends them to the agent application. The summaries appear within a few seconds after an interaction ends. Agents can modify the summaries if they want to change the text or add details. The agent must save the summary to complete the interaction.

AutoSummary works with CXone Agent, CXone Agent Integrated, CXone Agent Embedded, and CXone Agent for Microsoft Teams.

Key Facts about AutoSummary

Enlighten AutoSummary: