The challenge
High-performing CSMs were setting an unspoken standard that the rest of the team had no visibility into and no structured path to reach. Coaching was manager-dependent, inconsistent across the team, and subject to unconscious bias. The best CSMs were carrying institutional knowledge that no one else could access. Leadership suspected a performance gap but had no consistent baseline to measure against and no scalable way to close it.

​

The solution
An evidence-based performance baseline was established from industry best practices and senior CSM interactions then validated  before any AI was involved. A custom AI coaching agent was then built and trained on that validated baseline, not generic CS frameworks or unreviewed data. Every output was evaluated against human judgment before deployment. The result was consistent, bias-free coaching available to every CSM — with skill development no longer dependent on manager bandwidth, tenure, or personal style.

​

The origin

Leadership wanted to uplevel their CSMs but having every manager shadow calls was not feasible — and it introduced bias. CSMs would invite managers to their best calls, not their typical ones. When asked what the biggest skill gaps were across the team, leadership did not have that answer. There was no baseline and no consistent view of what good looked like.

The rubric came first. Working directly with senior ECSMs to define best-in-class performance, then validating those standards against industry benchmarks, created a practitioner-grounded baseline. Once that standard existed, the coaching agent could be trained against something real — not a generic framework, but an evidence-based standard that reflected what enterprise CS excellence actually looks like.

​

View a template report below. Note: Information is generic to protect proprietary information