Sales coaching with AI scorecards (L3)
Gong / Chorus AI scores every call against your rubric. Managers coach to the gap, not to vibes.
The steps
- 01
Build a 5-point rubric
The foundation of AI coaching is a focused, objective rubric. Start by identifying 4-6 high-impact behaviors (e.g., Discovery Depth, Objection Handling, Next Steps) that correlate with closed-won deals. • Open Gong or Chorus and navigate to the 'Scorecards' or 'Coaching' tab. • Create a new Scorecard titled 'Core Sales Execution'. • For each criterion, write a clear 'Evidence-based' description. For example, for 'Pain Discovery,' don't just say 'Find pain'; write 'Did the rep ask at least two follow-up questions to understand the business impact of the prospect’s problem?' • Assign a 1-5 scale or simple Yes/No toggles. • Owner: Sales Enablement or Head of Sales. • Time Estimate: 2-3 hours for workshop and setup. • Pitfall: Adding too many criteria (over 6) which confuses the AI and the rep. • Definition of Done: A published scorecard in your CI tool with 5 clearly defined, measurable criteria.
- 02
Configure AI-driven scoring
Now, map your rubric to AI-driven tracking so the software can 'read' the transcript and score it automatically. • In Gong, go to 'Settings' > 'AI Scorecard' or 'Trackers'. In Chorus, look for 'Smart Playlists'. • Link each rubric criteria to specific 'Trackers' or keywords. For 'Next Steps,' configure the AI to search for phrases like 'calendar,' 'next Thursday,' or 'follow up meeting.' • Set a threshold for 'Passing'. For example, if the AI detects the competitor name and a rebuttal, give full points for 'Competitive Positioning.' • Owner: RevOps or Sales Ops. • Time Estimate: 3-4 hours to tune the AI model. • Pitfall: Not testing the trackers. Run the AI against 10 old calls to ensure it isn't flagging 'false positives.' • Definition of Done: AI Scorecards are actively running and generating scores for 100% of recorded calls.
- 03
Identify coaching outliers
Instead of managers wasting hours listening to random calls, use AI to filter for the 'teachable moments.' • Navigate to the Coaching Dashboard in your CI tool. • Filter for calls in the bottom 20% of scores (red flags) to find coaching gaps, and the top 10% (gold standard) to find snippets for the team library. • Create a Saved View for managers called 'Coaching Focus' that isolates calls with low 'Discovery' scores but high 'Talk Ratio.' • Owner: Sales Manager. • Time Estimate: 30 minutes per week. • Pitfall: Getting stuck in the 'middle.' Only reviewing average calls leads to average results; focus on the outliers. • Definition of Done: Managers spend their first 15 minutes of the week reviewing the outlier calls identified by the AI scorecard.
- 04
Execute gap-based coaching
Move from 'vibes-based' coaching to data-backed plans. Don't coach them on the last call they had; coach them on their aggregate trend over 20 calls. • Open the Rep Performance view. Look for a specific rubric item where the rep consistently scores below 60% (e.g., 'Pricing Mentioned'). • Write a 30-day coaching plan: 'Rep will focus on introducing the price before the 40-minute mark.' • Use the 'Comment' or 'Tag' feature in Gong to leave feedback directly on the specific timestamps flagged by the AI. • Owner: Sales Manager. • Time Estimate: 1 hour per month per rep. • Pitfall: Giving the same feedback to everyone. Use the AI delta to personalize the coaching to that specific rep’s weakness. • Definition of Done: Every rep has one specific AI-identified skill gap they are working to improve this month.
- 05
Track the improvement curve
The final step is proving that coaching is working by looking at the improvement curve. • Go to the 'Coaching Analytics' or 'Insights' tab. • Export a report showing the 'Adherence Score' for your rubric over a 90-day period. • Cross-reference this with CRM data (Salesforce/HubSpot) to see if reps with improved AI scores are also seeing a higher 'Stage 2 to Close' conversion rate. • Build a 'Coaching Leaderboard' showing which managers have the highest lift in their team's AI scores. • Owner: RevOps. • Time Estimate: 2 hours per quarter. • Pitfall: Assuming a high AI score equals a high quota attainment without checking CRM data. • Definition of Done: A quarterly review deck showing the correlation between AI scorecard improvements and revenue growth.
Next playbooks
Unfamiliar terms are defined in the AI and Revenue Dictionary. Related frameworks live in the framework library.