AI-generated summary
into unified configurations that enable consistent interaction analysis at scale.
Core program capabilities include topic detection that identifies business-level intents in interaction transcripts, sentiment analysis that evaluates customer emotional trends and experience risks, AI-powered summaries and insights that provide visibility into outcomes and customer intent, and automated agent scoring that generates evaluations to support quality management and coaching. Programs support multiple dialects and languages to accommodate different business intents across contact center regions. Topics must be defined before program addition but can be edited afterward, with the program applying the topic version active at publish time. A preconfigured default program with predefined topics is included automatically.
Access to program functions requires specific permissions, with different administrative roles having distinct capabilities. Quality administrators manage queues, flows, and scoring but cannot modify speech and text analytics settings, while speech and text analytics administrators have full access to all program settings. In Genesys Cloud EX, programs can only be assigned to queues, as flows are not supported in that environment.
System limitations include support for up to 500 topics per dialect with a maximum of 15 selectable dialects, and a maximum of 500 unpublished topics. Programs with more than 500 topics experience performance indicators during editing or deletion, with certain features becoming unavailable. Agent scoring supports up to 20 enabled rules per program. Organizations can create a maximum of 100 programs, with each program limited to 500 queues and 500 flows; the entire system supports a maximum of 20,000 queues and flows combined.
Program configuration requires mandatory fields including a unique, case-sensitive name with a maximum of Programs support up to 20 tags per program and topic and must be saved as drafts before publishing. Publication triggers processing of all new interactions according to the program's configuration. Users can publish up to 10 programs simultaneously, with automatic force publishing applied to any topics containing unpublished changes.
Sentiment and empathy analysis features determine whether customer sentiment scores and agent empathy scores are generated for program interactions. Both settings are enabled by default but can be disabled at either the program or organization level. Organization-level settings override program-level configurations; when organization settings are disabled, analysis does not occur regardless of program settings, though program-level values are preserved for future re-enablement. Analysis states display as Enabled, Disabled (Program-level), or Disabled (Org-level) to indicate the source of any restrictions.
Automated agent scoring leverages AI-driven auto-complete evaluations configured through scoring rules, allowing administrators to specify interaction volume, target agents, and select published agent auto-evaluation forms. The system automatically assigns the last agent to participate in a conversation as the evaluated agent. For digital interactions, automated evaluations are regenerated each time a new response is added to the thread, ensuring scores reflect the most current transcript context. Program copying functionality allows duplication of existing configurations, with copied programs existing only as drafts until saved. Programs are accessed and managed through the Menu > Conversation Intelligence > Speech and Text Analytics > Programs interface, which displays available programs with creation, publishing, and deletion capabilities, along with filtering options by tags, queues, flows, and search functionality