Your complete guide
From first click
to your score in 15 minutes.
Six steps. One number that tells you exactly what your leadership is costing — and a 90-day path to close the gap.
The six steps
Answer 7 context questions
Role, org size, years in leadership, compensation range, and two behavioral context questions. These calibrate your score to your specific situation — a VP at a 500-person company has a different drag ceiling than a CEO at a 50-person startup. Takes about 2 minutes.
score.html → intake wizardCopy your diagnostic prompt
After the intake, you'll receive a prompt tailored to your dominant pillar. Copy it — one click. You're about to take it into your AI tool.
score.html → prompt copyWhich AI tool to use
Run the prompt in your AI tool — then come back
This is the only moment where you leave the app. Paste the prompt into M365 Copilot (or your AI tool), let it run — usually 30–60 seconds — then copy the output block it produces. Come back and paste it on the Score page.
~5–8 minutes · in your AI toolWhy does this step exist?
Your AI tool has something we don't: direct access to your behavioral signals —
your actual calendar, the timestamps on your emails, which meetings you attended, how decisions
routed across your team. That data lives in your Microsoft or Google tenant. It never leaves it.
We designed the diagnostic this way on purpose. Your behavioral data stays with you.
We only receive the anonymous, structured output your AI tool produces — numerical scores, not content.
Think of it as a blood test: your doctor reads the numbers, not the blood itself.
Without your AI signals
- Self-reported only — subject to blind spots
- Score shown as a ±15 point range
- No behavioral evidence to anchor the calibration
- Harder to track true movement at Day 21/45/90
With your AI signals
- Behavioral evidence — patterns observed, not estimated
- Point estimate score — single number, tighter range
- Calibration anchored to real data
- Pulse re-scores at Day 21/45/90 show true movement
Paste the output — receive your score
Back on the Score page, paste your AI output into the signal block field and click score. In seconds you'll see: your STIx composite score (25–125), five pillar breakdowns, your annual leadership drag cost in dollars, and your dominant pattern. This is your Day 0 baseline. It stays locked as your reference point for 90 days.
score.html → paste + scoreLock in your behavioral rule
Your report surfaces a single behavioral rule derived from your dominant pillar — a Stop/Start/Because commitment you can hold for 90 days. You can edit it, make it specific to your situation, then lock it in. This is the only thing you're committing to track. One rule. Not five. The research is clear that single-focus behavioral change outperforms multi-target interventions by a factor of 3x over a 90-day window.
score.html → rule lockRun the 90-day protocol
Your dashboard activates. Three milestones: Day 21 (first pulse re-score), Day 45 (mid-protocol check), Day 90 (full composite re-score). Weekly Friday check-ins ask three pillar-specific questions and whether your rule held. At Day 90 you see your total score delta — the measurable distance between where you started and where the pattern is now.
dashboard.html · 90 daysWhat the data shows
Leaders in the founding cohort who completed the full protocol reduced their dominant pillar score by 12–25 points. That maps to 3–5 hours per week recovered — documented over 90 days, not estimated.
Why you can trust the score
Your score is not produced by AI alone — and it's not a self-report questionnaire. It's a calibrated composite. Here's how it splits.
75%
Calibrated against peer-reviewed research
The core score is calculated by our proprietary engine — not by your AI tool.
It normalizes your behavioral signals against calibration drawn from external peer-reviewed behavioral research
(Bain, Cross & Grant, Perlow, and others) combined with the author's original diagnostic cohort of leaders,
then adjusts for role complexity, organizational transmission (how far the pattern has spread to your team),
and pattern entrenchment. The calibration reflects independent published research, not AI inference.
Why 75% and not 100%? Pure formula without behavioral evidence
produces the same score for two leaders whose patterns look identical on paper
but behave very differently in practice. The AI signal anchors the formula to
what's actually happening in your calendar and inbox.
25%
AI signal anchor
Your AI tool's behavioral read of your signals contributes 25% of the final score —
enough to anchor the calibration to your real data without letting AI inference dominate.
Why not more? AI models have no access to the cohort benchmarks,
role-specific calibration, or transmission research. They observe the signal;
the formula interprets it.
Research the formula draws on
Cross, Rebele & Grant · Harvard Business Review · 2016
"Collaborative Overload" — knowledge workers spend 85%+ of their week in collaborative work; senior leaders report 2–3 hours per day in non-strategic meetings. The drag benchmarks in the formula are calibrated to the upper bound of this documented range.
Perlow, Hadley & Eun · Harvard Business Review · 2017
"Stop the Meeting Madness" — 65% of senior managers report meetings prevent them from completing their own work; 71% say meetings are unproductive and inefficient. Meeting Theater pillar thresholds are set against this documented baseline.
Hackman · "Leading Teams" · 2002
Team dysfunction follows a compounding model when two or more failure modes are active simultaneously — single-failure-mode intervention is insufficient. This is the basis for the co-occurrence pattern amplifiers in the scoring engine.
Schein · "Organizational Culture and Leadership" · 1985
Culture is primarily defined by what the leader pays attention to, measures, and reacts to. After 5–7 years, the leader's behavioral patterns become the culture — they persist even after the leader changes behavior. This is why tenure and org size factor into the organizational transmission calculation.
Donaldson & Grant-Vallone · 2002
Cohen's κ for behavioral inference from calendar data averages 0.60–0.70 — meaning calendar-based behavioral signals have strong but not perfect inter-rater reliability. The scoring engine accounts for this uncertainty in its confidence weighting, which is why AI-only scores would overstate precision.