We had a similar challenge when scoring more subjective behaviors like empathy, especially around distinguishing opportunity vs execution.
In one of our implementations, we explicitly separated:
-
situations where the behavior was expected and delivered
-
situations where it was expected but missing
-
and cases where there was no opportunity (neutral)
This helped avoid penalizing agents when empathy simply wasn't applicable in the interaction.
We shared a real-world example of this approach in a high-volume operation here:
https://community.genesys.com/discussion/transforming-quality-management-with-ai-scoring-in-a-high-volume-operation#bm655c76a4-924f-465b-b318-eb25cf4aea28
Would be interested to hear how others are handling similar scenarios as well.
------------------------------
Mateus Nunes
Tech Leader Of CX at Solve4ME
Brazil
------------------------------
Original Message:
Sent: 01-27-2026 17:14
From: Doug Couch
Subject: Quality: Automated Evaluation Forms
I currently use AI scoring for things such as Scripting and Patient Experience. For example, I have it scoring for professional and courteous language on the call. I think the most unique question I have is related to empathy. I have three outcomes possible:
1) Transcript indicates a situation that required an empathetic response was identified and an empathetic response was provided.
2) Transcript indicates a situation that required an empathetic response was identified but no empathetic response was provided.
3) Transcript did not indicate a situation required an empathetic response.
I was wondering how others were using the AI Scoring feature. Any unique applications of the evaluation form? Any unique questions you've created?
#General
------------------------------
Doug Couch
Quality Assurance Manager
------------------------------