Post-Coaching Call AI Phase 2
Executive Summary
Every coaching call at RRI is a one-time interaction. The client has a session, maybe takes notes, and the insights fade within days. The coach moves to the next client. No institutional memory. No branded follow-up. No data accumulation.
Post-Coaching Call AI changes this. It records every coaching Zoom call, transcribes it through Tony’s coaching standards, and emails a branded recap to the client — written as if from Tony. Action items are extracted automatically. The next session prep is generated before the client even logs in.
This saves $364K/year compared to the Salesforce Einstein approach that was previously evaluated. But the cost savings are secondary. The real value is that every coaching session becomes a branded touchpoint, and Tony AI accumulates real coaching data from actual sessions — making Tony AI smarter with every call.
Jay Lane’s team has already proven the concept with Phase 1 coaching AI tools. Phase 2 extends this to full call recording, transcription, and branded delivery across Chris Schenke’s entire coaching fleet.
What Needs to Happen
- Build Zoom recording integration (webhook on call end) — Register a Zoom webhook that triggers when a coaching call ends. Automatically download the recording and queue it for processing. Handle consent and recording disclosure requirements per state/country.
- Transcription pipeline (Whisper or similar) — Process call recordings through a transcription model (Whisper, Deepgram, or AssemblyAI). Speaker diarization to separate coach and client voices. Output: timestamped, speaker-labeled transcript.
- Apply Tony’s coaching framework templates for analysis — Map transcript segments to Tony’s coaching frameworks: Six Human Needs, Triad (Focus/Language/Physiology), Pattern Recognition → Utilization → Creation. Identify breakthroughs, commitments, and areas of resistance.
- Generate branded email recap using Claude/GPT — Transform the analyzed transcript into a polished email recap in Tony’s voice. Include: session summary, key insights, action items with deadlines, and a motivational close. Brand it as a Tony Robbins coaching communication.
- Client follow-up automation (action items, next session prep) — Extract action items and create automated follow-up reminders. Generate next-session prep document for both coach and client. Track completion of commitments between sessions.
- Coach dashboard with session analytics — Build a dashboard for coaches showing: sessions conducted, client progress over time, framework utilization patterns, and outcome tracking. Help coaches identify which approaches work best for which client types.
- Tony AI accumulates real coaching data from sessions — Feed anonymized coaching patterns, successful interventions, and framework applications into Tony AI’s training data. Tony AI becomes a better coach because it learns from real sessions, not just Tony’s books.
- Roll out across coaching fleet (Chris Schenke’s team) — Deploy to all active coaches. Training session on how to use the dashboard and review AI-generated recaps before they send. Feedback loop for quality improvement.
Claude Code acceleration: The entire AI pipeline — Zoom webhook handler, transcription integration, framework analysis prompts, branded email generation, and follow-up automation — is pure code that Claude Code can generate rapidly. The coach dashboard is a standard data visualization project. Estimated savings: 2-3 weeks (from 6 weeks to 3-4 weeks).
Completion Criteria
- Coaching calls automatically recorded and transcribed with speaker diarization
- Branded email recaps generated and sent within 1 hour of call end
- Action items extracted with automated follow-up reminders
- Coach dashboard live with session analytics and client progress tracking
- Tony AI receiving anonymized coaching data to improve recommendations
- Deployed across Chris Schenke’s full coaching fleet
- $364K/year cost savings validated vs. Salesforce Einstein alternative
Initiative Attributes
Tools Required
| Tool | Purpose | Cost |
|---|---|---|
| Zoom API | Call recording webhooks, recording download, meeting metadata | Existing Zoom license |
| Whisper / Deepgram | Speech-to-text transcription with speaker diarization | ~$0.006/min (Whisper) or $0.0125/min (Deepgram) |
| Claude API | Framework analysis, branded recap generation, action item extraction | API inference costs (~$0.50-$2 per session) |
| Email service (SMTP2Go) | Branded coaching recap delivery | Existing infrastructure |
| Coach dashboard | Session analytics, client progress, framework utilization patterns | Build cost (included in timeline) |
Related Risks
No direct risk factors mapped to this initiative. S6 is a relatively self-contained AI pipeline project with Jay Lane as the proven owner. The primary risks are operational: ensuring recording consent compliance across jurisdictions, and coach adoption of the new workflow. Jay’s track record (30+ AI tools deployed in 180 days) de-risks execution significantly.