AI-generated summary
Genesys Cloud provides comprehensive quality management capabilities that enable teams to evaluate agent interactions through an advanced evaluation system featuring AI-assisted scoring with transparent justification. The platform allows evaluators to review recordings, emails, and complete questionnaires via an inbox or evaluator dashboard, with quality teams able to manage evaluation forms, create scoring indices, generate calibration exercises, and rescore interactions as needed.
Key functionality includes AI scoring for voice and digital interactions, limited to 50 evaluations per agent daily, flexible evaluation scheduling up to 45 days in advance, and evaluation assistance through topic and phrase detection. The system supports special question types including critical, fatal, and conditional questions, while providing single-click access to AI-suggested answer explanations. Access requires Quality Administrator or Quality Evaluator roles, with additional permissions necessary for AI scoring features.
Technical capabilities encompass interaction playback with secure pause indicators, phrase matching for configured topics, and transcript navigation. When AI generates evaluation answers, an AI icon appears next to the respective question, and the system displays the reasoning behind the AI-suggested answer. Users with appropriate permissions can access the reason for selection, which can be copied to the comment section. In cases where AI scoring misaligns with the reason for selection, manual editing is available; alignment can be improved through AI model enhancement or accuracy rating improvements via prompt filters in automated scoring options.
The system supports secure pause functionality to temporarily halt recordings and exclude sensitive information such as credit card numbers, with audible beeps indicating omitted segments. When evaluation assistance conditions identify multiple phrases for configured topics, they appear under the answered question; clicking a phrase navigates the audio playback and transcription to that point in the interaction. Evaluations can function as scoring indices for calibration exercises, and interactions can be rescored using the Re-score function.
Notable limitations include restrictions on multimedia content in messaging transcripts, a requirement for quality evaluator assignment for evaluation completion, and current formatting issues with dates and times in Arabic dialect views. AI scoring requires specific permissions including Quality Evaluation View Sensitive Data and Speech and Text Analytics Data View Recording access.