Defining Custom Evaluation Form Scoring Logic and Fatal Questions

How-To-Articles
1

Defining Custom Evaluation Form Scoring Logic and Fatal Questions

Executive Summary & Architectural Context

Quality Management (QM) in a contact center is more complex than a simple 1-to-100 grade. A Quality Evaluation Form is a legal and compliance document. If an agent is perfectly polite, solves the customer’s technical issue rapidly, and closes the ticket (scoring 95/100 on soft skills), but fails to verify the caller’s identity (a HIPAA/PCI violation), that interaction must not receive a passing grade.

The architectural mechanism to enforce compliance within Genesys Cloud QM is the Fatal Question. By engineering specific scoring weights and critical failure conditions, you guarantee that compliance breaches immediately override standard performance metrics, resulting in an automatic failure for the entire evaluation.

This masterclass details how to construct non-linear scoring logic, apply Fatal criteria, and handle “Not Applicable” (N/A) impacts on the mathematical denominator.

Prerequisites, Roles & Licensing

  • Licensing: Genesys Cloud CX 2 or 3 (WEM Quality Add-on).
  • Roles & Permissions: Quality > Evaluation Form > Add/Edit.

The Implementation Deep-Dive

1. The Anatomy of Form Weighting

By default, Genesys Cloud distributes points evenly. If you have 10 questions, each is worth 10%. This is an anti-pattern.

  1. Navigate to Admin > Quality > Evaluation Forms. Create a new form.
  2. Group your questions logically: Group 1: Soft Skills, Group 2: Technical Accuracy, Group 3: Compliance.
  3. Select the Group Weighting toggle.
  4. Assign weights: Soft Skills = 20%, Technical Accuracy = 50%, Compliance = 30%.
  5. Logic: This ensures that failing to use the customer’s name (Soft Skill) mathematically damages the score far less than providing the wrong technical answer.

2. Engineering the “Fatal” Question

A Fatal question destroys the total score regardless of performance in other categories.

  1. In the Compliance group, add a Multiple Choice question: “Did the agent authenticate the caller using two factors before accessing the account?”
  2. Add answers: Yes and No.
  3. On the No answer, toggle the Fatal checkbox.
  4. Scope of Fatality: Genesys Cloud provides two levels of fatality:
    • Fatal to Group: If checked, the agent scores a 0% for the Compliance group, but keeps the points they earned in Soft Skills.
    • Fatal to Evaluation: If checked, the entire evaluation immediately hard-fails, resulting in a final total score of 0%, regardless of the math in other groups.
  5. Best Practice: For strict security protocols (PCI, HIPAA), always use Fatal to Evaluation.

3. Handling the N/A (Not Applicable) Denominator

One of the most mathematically disastrous configuration errors is misconfiguring the N/A option.

If a form has 100 possible points, and Question 5 is worth 10 points: “Did the agent offer a cross-sell product?”

  • If the call was a complaint, offering a cross-sell is inappropriate. The evaluator selects N/A.
  • The Trap: Does N/A mean 0 points awarded out of 100? If so, the agent’s maximum possible score is artificially capped at 90%, penalizing them for a scenario outside their control.
  • The Fix: Genesys Cloud natively handles N/A dynamically. When an evaluator selects N/A, the system automatically removes those 10 points from the total possible denominator. The agent is now graded out of 90 total points, meaning they can still achieve a 100% final score. Ensure the N/A checkbox is toggled on questions that are situationally dependent.

Validation, Edge Cases & Troubleshooting

Edge Case 1: Form Versioning and Calibration History

If you publish an Evaluation Form, and 3 months later Legal demands you add a new Fatal question regarding a new disclosure, you cannot edit the live form.

  • Troubleshooting: You must create a new Version of the form.
  • The Impact: When you view historical analytics, scores generated under v1 of the form are mathematically distinct from scores generated under v2. If you are comparing Agent A (evaluated mostly on v1) against Agent B (evaluated on v2 with the new Fatal question), the comparison is invalid. BI teams must filter Quality reports by Form Version to ensure fair performance reviews.

Edge Case 2: AI Auto-Scoring Conflicts

If you are using Genesys Cloud’s Automated Quality Management (AQM), the AI evaluates 100% of interactions based on Speech Analytics topics.

  • The Trap: The AI is currently limited in its ability to detect “silence” or “omission” with 100% certainty (e.g., proving the agent didn’t authenticate the caller).
  • Solution: Do not map AQM AI logic directly to Fatal to Evaluation questions without human oversight. If the AI hallucinates or the transcription engine misses the authentication phrase, the agent gets an automatic 0% and their bonus is destroyed. AI should score standard metrics; human evaluators should audit Fatal compliance questions.

Official References