Quality monitoring (QM) is the process of listening back to calls (via recording) to score quality according to criteria, then use those results to coach agents. The goal is not to "spot errors" but to improve steadily: the agent knows where he is doing well/not well, and the team has data for training. Zenify credits AI support with a 35% reduction in processing time and 98.4% accuracy.
TL;DR
- QM = score calls according to criteria + coach from there.
- Mark according to: tone, process, information, results.
- Don't "look for errors", the goal is to improve.
- Score a random sample, evenly for each agent, each week.
- Coaching based on specific scores, not general.
What is Quality monitoring?
QM includes loop:
Grading: listen to the call recording, grade according to the criteria table.
Synthesis: each agent's score, common error groups, highlights.
Coaching: guides agents to correct errors and maintain strengths.
Re-measure: After 2-4 weeks see if the score increases.
Different from QA: QM often focuses on interaction quality (response); QA more broadly includes process and compliance (see What is QA).
Scoring according to what criteria?
Group | Example criteria
Accent & attitude | Greet in a friendly, non-irritating tone
Process | Correct processing steps, confirmation of information, recording
Information | Answer correctly, sufficiently, not wrongly committed
Results | Problem solved, customer satisfied
Compliance | Don't make promises beyond your authority, follow the rules
Each criterion is scored on a scale (0/1, or 1-5), the total score is a percentage or average (see QA score).
Effective way to grade
1. Score each agent randomly and evenly
Each agent makes 2-5 calls/week, enough representation, no effort.
Randomly avoid "easy scoring".
2. Dial quickly after calling
Dot close to the call time, the agent still remembers the context.
Take notes on specific examples (statements, conversations) when marking.
3. Calibration by multiple people
Two people judge the same call → compare → agree on criteria.
Reduce grading differences between managers (see Calibration QA).
4. Group analysis
Team-wide errors (wrong procedures, lack of greetings) → group training.
Personal errors → private coaching.
Coaching from QM results
Say points + specific examples: "Yesterday's call to say hello lacked information confirmation."
Ask the agent to self-assess: "Do you think it was handled well that time?", so that the agent can realize for himself.
Choose 1-2 improvement points each week: don't accumulate 10 mistakes at once.
Practice: role play a good or bad situation.
Re-measure after 2 weeks: see if there is progress or not.
Golden rule: Coaching criticizes errors, does not criticize people, avoids agents being afraid of being graded.
Common QM errors
Only dot to "check": agent avoids and hides calls → true quality is not revealed.
Uneven grading: only a few agents are graded → unfair.
No specific examples: general scores → agent doesn't know what to fix.
Not tracking improvements: After marking, leaving it there → wasting effort in vain.