Implementing Automated Sentiment-Based Agent Coaching Assignment in Quality Management

How-To-Articles
1

Implementing Automated Sentiment-Based Agent Coaching Assignment in Quality Management

What This Guide Covers

  • Automating the Quality Management (QM) lifecycle by using real-time sentiment analysis to trigger targeted coaching sessions for agents after high-friction interactions.
  • Configuring Genesys Cloud Policy Management and the Workforce Management (WFM) API to automatically schedule development time for agents who encounter consecutive negative sentiment interactions.
  • The end result is a data-driven coaching culture that proactively addresses agent performance gaps and reduces turnover by providing support when it’s most needed.

Prerequisites, Roles & Licensing

  • Licensing: Genesys Cloud CX 3 or AI Experience add-on (for sentiment) + WEM (for Coaching/WFM).
  • Permissions: Quality > Policy > Add, WFM > Schedule > Edit, Quality > Coaching > Add.
  • Infrastructure: Interaction Analysis must be enabled for the relevant queues.

The Implementation Deep-Dive

1. Defining the “Coaching Trigger” Policy

The first step is to identify which interactions deserve an automated coaching session. You don’t want to coach every single negative call (some customers are just difficult), but rather patterns of failure.

Implementation Steps:

  1. Navigate to Admin > Quality > Policies.
  2. Create a new policy with a matching criteria: Sentiment Score < -0.6.
  3. Add a second criteria: Agent Sentiment Score < -0.3. This ensures we are catching interactions where the agent’s behavior may have contributed to the friction.
  4. Set the Action to “Flag for Coaching.”

Architectural Reasoning:
By combining customer and agent sentiment, you filter out “unavoidable” negative calls (where the agent was professional but the customer was angry) and focus on “addressable” coaching opportunities where the agent’s tone also shifted negatively.

2. Automating Coaching Appointment Creation

Once a policy flags an interaction, you can use a Genesys Cloud Data Action or a background script to create the actual coaching appointment.

Implementation Steps:

  1. Use the Notification API to listen for v2.quality.evaluations or custom policy triggers.
  2. When a trigger is received, call the Coaching API (POST /api/v2/quality/coaching/appointments).
  3. Set the attendeeIds to the agent and their direct supervisor.
  4. Link the specific conversationId to the appointment so the supervisor can review the transcript during the session.

The Trap:
Creating too many appointments. If an agent has a bad day, you don’t want 10 coaching sessions appearing on their calendar. Implement a Rate Limiter in your automation logic (e.g., “Max 1 automated coaching session per agent per week”).

3. Integrating with WFM for Automatic Scheduling

A coaching appointment in Genesys Cloud doesn’t automatically “carve out” time on the agent’s WFM schedule. Without WFM integration, the agent will be expected to take calls during their coaching session.

Architectural Reasoning:
Use the Workforce Management API to synchronize the coaching appointment with the agent’s official schedule.

  1. Fetch the agent’s current schedule using GET /api/v2/workforcemanagement/businessunits/{id}/weeks/{week}/schedules.
  2. Find a “Low Occupancy” window (e.g., during a period of predicted low call volume).
  3. Use the POST /api/v2/workforcemanagement/businessunits/{id}/managementunits/{muId}/schedules/activities endpoint to insert a “Coaching” activity block that matches the coaching appointment time.

The Trap:
Scheduling coaching during peak call volume. This will hurt your Service Level (SL). Always use the WFM Intraday Monitoring stats to find the “Optimal” time for off-queue activities.

Validation, Edge Cases & Troubleshooting

Edge Case 1: Supervisor Availability Conflicts

  • The Failure Condition: An automated coaching session is scheduled, but the supervisor is already in a meeting.
  • The Root Cause: The automation only checked the agent’s schedule, not the supervisor’s.
  • The Solution: Integrate with the Microsoft Graph API or Google Calendar API to verify the supervisor’s availability before finalizing the coaching appointment in Genesys Cloud.

Edge Case 2: Agent “Coaching Fatigue”

  • The Failure Condition: Agent performance drops further because they feel “policed” by the automated system.
  • The Root Cause: High-frequency, low-context coaching triggers.
  • The Solution: Implement a Positive Reinforcement trigger. Create a secondary policy for “Highly Positive” sentiment interactions (Score > 0.8) and trigger an automated “Recognition” email or a “Kudos” badge in the Gamification module. This balances the “Performance Correction” with “Performance Celebration.”

Official References