Genesys Cloud - Main

How are you turning Empathy/Sentiment insights into coaching actions-especially when you're not reviewing interactions in real time?

Glenda,

Another great question! I would love to hear what the community has to say!

Cheers,

For us, the key is turning sentiment into a management indicator, not just a quality observation.

In our CX team, we have a report where we correlate Sentiment x NPS and then segment the results into quartiles. This allows us to identify which groups are performing well, which ones need attention, and where coaching or training actions should be prioritized.

Even without reviewing interactions in real time, this gives us a clear view of patterns and helps us act based on data instead of isolated cases. It becomes a practical way to connect customer perception, emotional signals, and agent development actions.

In short, it is coaching and operational management driven by data.

We stopped treating sentiment as a QA checkbox and started using it as a triage layer. In Genesys Cloud, we pull the Agent Sentiment Detail report weekly and filter by agent to catch anyone trending down over two weeks, rather than reacting to one bad day.

From there, we use the interaction transcript view to find the exact moment sentiment shifted instead of reviewing the full call, which saves a ton of time. We also cross-reference with the Agent Performance Summary, so when we sit down for a coaching session, we are looking at sentiment trend alongside CSAT and FCR together, not in isolation.

That combination makes the coaching conversation way harder to dismiss because the agent can see how their emotional tone is connecting to actual customer outcomes. The whole workflow lives in Genesys Cloud natively, so there is no data wrangling; it just becomes part of your regular coaching rhythm.

------------------------------
Chris Rodriguez
Contact Center System Administrator
------------------------------

Hi Glenda,

We've taken a slightly more operational angle alongside that. We initially used sentiment as a trigger, but found it could be quite noisy, so we shifted towards combining (and in some cases prioritising) topic-based signals.

This helps us surface specific interactions where:

• The customer was dissatisfied, and
• There's context (e.g. knowledge gap, repeat contact)

So instead of only looking at patterns, we're also enabling targeted, interaction-level coaching moments.

Longer term, we're also looking to use these flagged interactions as inputs for automated evaluations, so the same signals can drive both coaching and quality scoring.

I think both approaches complement each other well one for trend identification, the other for actionable coaching,