PUSH

Where the real leverage lives

Top performers are already top performers. A small percentage lift on their numbers is nice but it is not the growth story. The growth story lives in the middle seventy percent. These are the producers who are already on the team, already onboarded, already paid, already in front of clients every week. Moving a fraction of them toward the behaviors of the top fifteen percent is the largest unlocked revenue line on the income statement, and most firms never touch it because they do not know how. Push is the motion that touches it.

Continuous coaching, not annual training

The research on behavior change in skilled professions is clear. One-shot training events produce forgetting curves within days. What actually changes behavior is frequent, specific, personalized feedback attached to real work, delivered consistently over months. Push delivers exactly that. Every conversation a producer has gets scored against the codified behavioral model. Every score generates a small number of specific coaching prompts tied to the exact moves that need adjustment. The AI coaching agent surfaces those prompts in a cadence the producer can absorb without drowning, and manager dashboards show which producers are progressing, which are stuck, and which are ready for stretch assignments.

The numbers we see

Across our engagements, teams running Push typically see fifteen to forty-two percent improvement in overall conversation quality scores inside the first eight weeks. We have documented individual producers moving from a baseline score of zero to a score above ninety through deliberate AI practice reps combined with continuous coaching. The improvements are not subjective. They are measured against the same rubric the firm approved during Teach, graded by the same system that graded the baseline, and visible at the individual, team, and regional level.

What this motion feeds

Push does two things for the rest of the flywheel. First, it lifts the performance floor, which means the organization's current revenue run-rate goes up. Second, it generates an ongoing stream of scored data that shows which behaviors transfer easily, which resist change, and which turn out to matter less than Surface initially suggested. That feedback loop is what keeps the next motion — Compound — calibrated to reality rather than to a snapshot from six months ago.