Genesys Cloud October 2026 Release Highlights

With customer-first callbacks, Genesys Cloud dials the customer first and then connects an agent. Administrators can now choose whether that agent is reserved before the callback is placed or only after the customer answers.

Agents can now add a note to a shift trade request, such as the reason for the trade or the shifts they're willing to take in return. Administrators see the note when they approve or deny the request.

Attribute-based access control (ABAC) policies can now restrict transcript access based on a conversation's Customer Access Attribute. ABAC is checked before existing roles, permissions and divisions. If the policy denies access, the user can't view the transcript, even if their role would normally allow it.

This is the first of three ABAC releases, previously shared on the roadmap in VDS QBRs, that work together to support data separation across lines of business, regions or clients.

ABAC policies now control access to policy-based recordings and segments, snippet recordings and screen recordings. When ABAC allows access, existing permissions and division restrictions still apply.

ABAC now extends to quality management evaluations and surveys, with separate policies for each. A QA team might have broad access to evaluations across business units, while survey results for a specific client or region stay with the team that supports them.

Agents working in the ServiceNow Agent Workspace can now see a real-time transcript of voice calls in the active call viewer. Genesys Cloud sends each utterance to ServiceNow as the conversation happens.

Agents can now generate on-demand summaries of long or multi-agent email threads. Each summary covers customer context, key decisions, resolution status and follow-up actions, and agents can regenerate it as the thread continues.

While creating or editing a checklist, administrators can now test it against a selected or pasted transcript. They see which checklist items matched and which didn't, along with supporting evidence when it's available.

Analytics, Interaction Search and Evaluation views now include filters and labels that separate Agent Auto-Evaluation (Virtual Supervisor) evaluations from manual ones.