Agent coaching that actually works — the 3-30-90 model
Why quarterly coaching sessions rarely move the needle, and how a 3-30-90 rhythm with AI-scored calls changes agent performance in 3 months.
Most contact centers run coaching quarterly. Agent stats say the ratings barely change from one quarter to the next — so coaching gets blamed. In fact, the cadence is the problem, not the coaches.
Below is a model that works — the 3-30-90 rhythm — and how AI-scored calls make it possible.
The 3-30-90 rhythm
- 3 minutes of feedback within 24h of any high-signal call (great, terrible, first hard case)
- 30 minutes weekly 1-1 covering patterns and one skill to improve
- 90 days as the horizon for measurable behavior change
Each layer serves a different purpose:
- 3-minute feedback fixes muscle memory
- Weekly 1-1 builds identity as a specific kind of agent (“the one who’s great at empathy”)
- 90-day horizon lets change show up in the metrics
Why it fails without AI
Classical QA can’t feed the 3-minute layer. If reviewers sample 3% of calls, most agents have zero high-signal moments captured in a given week. The whole rhythm collapses to whatever the reviewer happened to catch.
With AI scoring every call:
- Great calls are surfaced automatically (all “excellent” scores → 3-min praise message)
- Terrible calls surface too (all “compliance risk” flags → same-day supervisor conversation)
- Weekly 1-1 has real data (not “here’s what I remember”)
What supervisors actually do differently
Before: 4 hours a week listening to random calls, hoping to find something useful.
After: 4 hours a week reviewing AI-scored highlights, running 3-minute check-ins, prepping weekly 1-1s from actual patterns.
The work shifts from finding signal to acting on signal.
Numbers you’ll see
Typical operations rolling out the 3-30-90 rhythm with AI-scored calls report:
- QA score up 12-18% in the first 90 days
- Agent tenure up 20-30% year-over-year (less churn)
- Time to competence for new agents down from 6-8 weeks to 3-4
What next
The rhythm is free. The AI is the multiplier that makes it feasible. Start by picking your top 3 QA criteria and running the 3-30-90 loop just on those. Once it’s stable, expand.
Want AI-scored calls feeding your coaching rhythm? Book an InOro demo.