Hi Nick, great questions. I've been testing Automated Agent Scoring using programs as well and wanted to share a few observations from hands-on usage.
The automated assignment and scoring flow itself is very solid and promising. The ability to score at scale without manual intervention is a big win.
That said, from a quality process perspective, a few gaps became very clear during testing:
• Distribution logic
Today the distribution feels very generic. It would be extremely valuable if automated scoring could better respect existing quality policies, especially in environments where not all forms are 100% AI-driven and still follow specific assignment rules.
• Release to agent (auto-feedback)
Evaluations scored by Virtual Supervisor are not released to the agent. This becomes a major limitation if the goal is auto-feedback and continuous improvement. If the evaluation is final enough to score performance, it should also support controlled release.
• Contestation workflow
There is currently no contestation path for AI-scored evaluations. This is especially critical for AI forms, where misinterpretations can happen. A human-in-the-loop contestation step feels essential to build trust and adoption.
Given these points, I wanted to ask the Genesys team:
Is there an action plan or roadmap to evolve Automated Agent Scoring into a more complete quality workflow, including release, contestation, and richer distribution logic?
Overall, the foundation is very strong. With these additions, Automated Agent Scoring could truly become a first-class quality process rather than just a scoring mechanism.
Thanks for opening the discussion.
------------------------------
Mateus Nunes
Tech Leader Of CX at Solve4ME
Brazil
------------------------------