Autonomous coaching nudges from call data (L5)
The system detects a coachable pattern across a rep's calls, drafts the nudge, and delivers it in Slack without waiting for a manager. Managers audit the nudges rather than write them.
The steps
- 01
Limit nudges to patterns, not single calls
Tool: Gong
A behavior must appear in at least three calls in two weeks before a nudge fires. Single-call feedback from a machine reads as surveillance. Owner: enablement. DoD: the threshold is configured and documented.
- 02
Deliver privately with an example attached
Tool: Slack
Direct message the rep with the pattern, two timestamps, and one suggested phrase. Never post in a public channel. Owner: enablement. Pitfall: public leaderboards of AI-detected weakness. It kills adoption in a week. DoD: all nudges land in direct messages with clip links.
- 03
Give the manager a weekly audit view
Tool: Manual
Managers see every nudge sent to their team and can flag a bad one. Bad nudges pause that pattern until it is retuned. Owner: sales managers. DoD: a weekly audit exists and at least one pattern has been paused or retuned.
- 04
Let reps dispute and record the dispute
Tool: Slack
A one-click dispute that a human reads. Disputes are the training signal for the rubric. Owner: enablement. DoD: dispute rate is tracked and reviewed monthly.
Tools in this playbook
Next playbooks
Unfamiliar terms are defined in the AI and Revenue Dictionary. Related frameworks live in the framework library.