Speech and text analytics is a set of features that uses natural language processing (NLP) to provide an automated analysis of an interaction’s content, to provide insight into customer-agent conversations. Speech and text analytics includes the transcription of voice interactions, analysis for customer sentiment and topic spotting, to create meaning from otherwise unstructured data.

Organizations can use this data to fulfill key use cases around agent performance improvement (for example, decrease AHT, increase FCR, sales conversion, and so on), compliance, customer satisfaction (for example, NPS), and customer business intelligence. 

Speech and text analytics analysis is performed against the interaction immediately after it is completed. However, if voice transcripts are needed with lower latency, it is possible to subscribe to transcripts through the Notifications API.  For more information, see the Genesys Cloud Developer Center.

For more information, see FAQs: Speech and text analytics .

Key features

• Voice transcription and digital transcripts – Get insight into conversations between external (customer) and internal (flows, agents, etc.) participants to see who is saying what. For voice interactions, audio is transcribed using our native transcription engine and the internal participant can be an IVR, voice bots, ACD, agent, conference or voicemail. For digital interactions (for example, email, message, or chat) the internal participant can be bots or agents.  This information can be used to improve employee training and feedback, and to identify business problems. For more information, see About voice transcription and About digital transcripts. 
• Customer sentiment analysis – Recognize a customer’s attitude during an interaction based on the language used during an interaction. By capturing the sentiment of the customer’s phrases, users gain valuable insight into the customer’s experience and subsequently use this information to improve service delivery. For more information, see About sentiment analysis.  
• Topic spotting – Detect topics of interest within conversations that are relevant to the business. For example, customer contact reasons, customer experience indicators, or expected agent behaviors based on a set of predefined phrases. A number of Out-of-the-box topics are provided.  For more information, see About programs, topics, and phrases.
• Interaction overview and details – A visual representation of the voice or digital interaction that enable agents and supervisors to review, recognize and determine customer sentiment and agent competency. For voice interactions the interaction overview includes a variety of controls (play, pause, annotate, live monitoring, adjust volume and speed, sentiment and topic markers, so on), For more information, see About the interaction overview and View an interaction’s conversation details. 
• Content search – Search for interactions based on the content of the interaction including words or phrases, customer sentiment and topics detected (coming soon). For more information, see Content Search view Content Search view.
• Analytics views – View aggregated data from speech and text in agent, queue and flow views. For more information, see About reports, views and dashboards.

Getting started

To get started with speech and text analytics, you must first enable voice transcription or select a default expected dialect for digital interactions. For more information, see the Getting started with speech and text analytics section in the Speech and text analytics article.