Building Custom Adherence Dashboards with Real-Time Exception Tracking

How-To-Articles
1

Building Custom Adherence Dashboards with Real-Time Exception Tracking

Executive Summary & Architectural Context

In a high-pressure contact center, a Real-Time Analyst (RTA) is the “Air Traffic Controller” of the floor. Their job is to ensure that the agents who are scheduled to be working are actually working. However, the standard WFM adherence interface is often a giant, flat list of 300+ names. When an agent goes “Out of Adherence”-for example, they are still on “Break” even though the schedule says they should have returned five minutes ago-the RTA has to manually click on the agent, look up their schedule, check their current status, and decide if a supervisor needs to be alerted. If 20 people go out of adherence during a shift change, the RTA is instantly overwhelmed. They spend their time scrolling through a list and miss the “Critical” failures (like a Tier 3 Senior Agent missing their shift entirely) while focusing on “Minor” ones (someone being two minutes late for a coffee break). The result is a understaffed floor and a crashing Service Level.

A Principal Architect replaces this “List-Based Monitoring” with a Priority-Driven Adherence Dashboard. By leveraging the WFM Real-Time Adherence API, you can build a custom view that ranks agents by the Severity and Duration of their non-adherence. Instead of a list of 300 names, the RTA sees the “Top 5 most impactful issues” at the very top of their screen.

This masterclass details how to architect a real-time adherence engine that prioritizes action and automates exception tracking.

Prerequisites, Roles & Licensing

Licensing & Permissions

  • Licensing Tier: Genesys Cloud CX 3 or WFM Add-on. NICE CXone WFM.
  • Granular Permissions:
    • WFM > Real-Time Adherence > View
    • WFM > Management Unit > View
    • Integrations > Action > Execute
  • Dependencies:
    • WebSocket Client: To receive real-time presence updates.
    • Middleware: (Python/Node.js) to aggregate API data into a priority-ranked dashboard.

The Implementation Deep-Dive

1. The Architectural Strategy: The “Impact Score”

Not all non-adherence is equal. A Principal Architect uses a Weighted Impact Score to rank the dashboard.

The Formula:
Impact_Score = (Duration_Out_of_Adherence_Seconds * Agent_Skill_Weight) / Service_Level_Sensitivity

The Priority Logic:

  1. Critical (Score 100+): Agent is scheduled for a high-value queue but is currently Offline or on an unapproved Break.
  2. Major (Score 50-99): Agent has been out of adherence for more than 10 minutes.
  3. Minor (Score < 50): Agent is 2 minutes late for a break.

2. Implementing the Real-Time Adherence API

The WFM API provides a stream of adherence data that you must “Pivot” for your dashboard.

The API Call:
GET /api/v2/workforcemanagement/managementunits/{muId}/adherence

The Normalized Data Structure:

{
  "agentId": "abc-123",
  "actualActivityCategory": "Break",
  "scheduledActivityCategory": "OnQueue",
  "adherenceState": "OUT_OF_ADHERENCE",
  "impactMinutes": 12,
  "priority": "HIGH" // Calculated by your middleware
}

3. “The Trap”: The “False Positive” Notification Storm

The Scenario: An agent’s shift ends at 5:00:00 PM. They click “Logout” at 5:00:05 PM.

The Catastrophe: Because of the 5-second “Sync Delay” between the Agent Desktop and the WFM Engine, the system flags the agent as “Out of Adherence” (Scheduled: Off-Queue, Actual: On-Queue). If your dashboard triggers an alert the millisecond someone goes out of sync, your RTA will be bombarded with hundreds of “False Alarms” every shift change. They will eventually start ignoring the dashboard entirely, rendering the whole system useless.

The Principal Architect’s Solution: The “Grace Period” Buffer

  1. The Delay Filter: Do not display any agent on the “Priority Dashboard” until they have been out of adherence for at least 120 seconds.
  2. The “Smart Sync” Logic: Before alerting, check if the agent’s status is “Pending Logout.” If so, suppress the alert.
  3. This “Debouncing” of the data ensures that the RTA only sees real, actionable issues, significantly reducing “Notification Fatigue.”

Advanced: One-Click “Instant Exceptions”

A Principal Architect integrates the “Tracking” with the “Fixing.”

Implementation Detail:
If the RTA sees that Agent Jane is out of adherence because she is in a “Manager 1-on-1” (verified via a Slack message), the dashboard should have a single button: “APPROVED EXCEPTION”.

  • When clicked, the middleware calls the WFM API to insert an “Administrative Exception” for that agent for the next 30 minutes.
  • This instantly moves the agent back into “In Adherence” status and updates the dashboard.
  • This saves the RTA from having to log into the main WFM UI, search for Jane, and manually edit her schedule.

Validation, Edge Cases & Troubleshooting

Edge Case 1: “Ghost” Sessions

The failure condition: An agent is showing as “On Queue” in the dashboard, but they are actually sitting in the breakroom.
The root cause: The agent’s browser crashed or they simply locked their screen without logging out.
The solution: Cross-reference Presence with Interaction Activity. If an agent is “On Queue” but has had 0 interactions and 0 minutes of talk time for the last 30 minutes, flag them as a “Stale Session” on the dashboard.

Edge Case 2: Multi-Skill “Partial” Adherence

The failure condition: An agent is scheduled for “Support” but they have manually switched themselves to “Sales.”
The root cause: The WFM engine sees them as “On Queue” (General) and marks them “In Adherence,” even though they are in the wrong queue for the current staffing need.
The solution: Use the Granular Adherence settings. Configure the dashboard to check the specific QueueID. If Scheduled_Queue != Actual_Queue, flag it as a priority mismatch.

Reporting & ROI Analysis

The success of an RTA dashboard is measured by Reaction Time.

Metrics to Monitor:

  • Average Out-of-Adherence Duration: Has the time agents spend “out of sync” decreased?
  • RTA Action Rate: Percentage of non-adherence events that were “Actioned” (Exception added or Agent contacted) within 5 minutes.
  • Service Level Recovery: Correlation between dashboard alerts and SL stabilization.

Target ROI: Expect to increase RTA efficiency by 50%, allowing a single analyst to manage double the number of agents with higher accuracy and significantly lower stress.

Official References