Methodology & Validation

Standards-stack foundation

Sextant is not a proprietary framework. It composes three industry-standard frameworks the buyer can independently verify and cross-reference.

Seven-pillar framework · source frameworks per pillar

Each pillar synthesizes the strongest vendor and academic frameworks for its dimension. Vendor names appear here for internal traceability but are paraphrased in client-facing copy to avoid jargon and avoid promoting the originating firm.

Pillars 6 and 7 are Sextant's differentiating pillars. Vendor frameworks underweight culture (Pillar 6) where the cohort actually fails; PE sponsors immediately recognize pilot purgatory (Pillar 7) as a board-legible metric.

Pillar 6 · nine validated psychometric instruments

Culture & Psychological Readiness is the most heavily-instrumented pillar in Sextant. Nine peer-reviewed instruments triangulate attitude, anxiety, trust, literacy (self-report and objective), behavioral intent, and change readiness. Items are administered verbatim from published validation papers to preserve norm comparability.

Plus three behavioral / typology layers — Microsoft 4-tier user typology (Skeptic / Novice / Explorer / Power User), resistance archetype (Champion / Early Adopter / Fence-sitter / Skeptic / Saboteur), shadow-AI disclosure — administered as multi-choice and scenario judgment rather than Likert.

Four analysis rules · pre-registered before fielding

1. Measure dispersion, not just averages. Every Pillar 6 sub-score reported as mean + variance + by-layer slice (C-suite / VP / Director / Manager / IC). High mean + high variance = pockets of fear hidden behind a healthy average.
2. Observe behaviors, not just attitudes. Self-report Likert pairs with multi-choice objective knowledge (AICOS), scenario judgment, and Sensor C telemetry when available.
3. Measure attitudinal AND structural signal at the management layer. Sensor A captures attitudinal (TPS-7, AIAS); Sensor B/C captures structural (budget authority, change-management time, sanctioned-tool access). "Binding constraint sits in middle management" claim comes from triangulation, not attitudinal data alone.
4. Cell-size collapse rule. N<5 layers collapsed with adjacent layer in fixed order (C-suite + VP → VP + Director → Director + Manager). Applied uniformly, never post-hoc.

Score scale · 0–4 maturity

All seven pillars use the same uniform 0–4 scale. Anchors are named, not numeric — buyers understand "you're at ad hoc on Data" more clearly than "you're at 1.4 out of 5."

Continuous interpolation between integer anchors is allowed (e.g., 2.7 = "managed approaching optimized"). Scale consistent with NIST AI RMF maturity language and Phoenix backend evaluation primitives.

Measurement methodology · how Sextant computes ROI

Every dollar value, percentage, and classification a CEO sees in the Sextant Command Center is produced by a documented formula — never editorial copy. This section is the audit trail. Five computations drive the entire surface: Drift Profile cost math, the Σ-Score composite, the Quantum opportunity envelope, the Reallocation Index (with its headcount + cultural risk classification rules), and the Agent Architecture savings model.

5.1 · Drift Profile · loaded-labor + tooling cost math

Say it"We add up what a recurring job costs you today — people, hours, software — then what it costs once the right AI tier handles the routine parts. The gap is recoverable drift, in dollars you can take to a board."

Per-workflow cost decomposition. Drift = the gap between what a recurring SOP costs today (loaded labor + tooling + other) and what the same SOP costs after Sextant routes the right agent tier to each subtask. Computed independently for every workflow, then rolled up to the Annual rollup line in the Drift Profile table.

Worked example — Helios "Monthly financial close": 12 × 80 × 3 × $325 + $90K tools = $1.025M baseline; future 12 × 18 × 2 × $325 + $30K = $270K; savings = $755K (74% lower).

5.2 · Composite readiness (Σ-Score) · pillar-weighted maturity

Say it"Think of it as a weighted GPA across seven readiness subjects — and we weight the subjects that matter most in your industry. One number, 0 to 4, that moves as you commit to tracks."

The single 0–4 score a CEO sees on the Quantum chart and Σ-Score panel. Weights vary by industry — manufacturing weights Data and Tech higher; SaaS weights Talent and Culture higher; finserv weights Risk higher. Weights sum to 1.0 and are published per-profile in the data layer.

Worked: Helios industry weights (mfg) → Strategy 0.15, Data 0.10, Tech 0.10, Risk 0.15, Talent 0.15, Culture 0.20, Value 0.15. Current composite = 1.44; with all 6 tracks selected → 3.15.

5.3 · Quantum opportunity envelope · ceiling, floor, selection cone

Say it"The band is your full range of outcomes — do nothing and you slide down the floor; commit fully and you reach the ceiling. The more you commit, the tighter and more certain the path becomes."

The band on the Quantum chart shows the full range of outcomes from no-action floor (declining) to all-tracks ceiling. The user's live selection line moves through that band; the selection cone tightens as commitment increases (more tracks selected = lower variance).

Worked: Helios ceiling = 3.15, no-action 18-month floor = 1.28, selection at 6/6 tracks → cone narrows 65% (uncertainty band ±0.35 → ±0.12 by +18mo).

5.4 · Reallocation Index · headcount + cultural-risk classification

Say it"This is the layoffs question, answered before they ask it. The index is the share of freed-up capacity we redeploy to higher-value work instead of cutting — higher means more growth and lower cultural risk."

The 0–100 score that headlines the Reallocation Map. Sextant treats AI as augmentation, not reduction: the Reallocation Index = percentage of reclaimed capacity that gets redeployed (to higher-value work, training, new initiative, or reserve) rather than eliminated. Pairs with two classification metrics that drive the Index narrative line.

Worked: Helios reallocation = 8% eliminated, 23% training, 41% higher-value, 15% new-initiative, 13% reserve. Index = 92. Net headcount = "stable" (elim ≤ 15%). Cultural risk = "minimal" (elim ≤ 10% AND training ≥ 15%).

5.5 · Agent Architecture savings · tiered-mix vs. all-frontier

Say it"Same output, a fraction of the cost. We put the frontier model only where judgment matters, let mid-tier handle delegation, and run value-tier autonomously overnight. The tiering is where the savings live."

Per-workflow cost savings from routing the right model tier to each subtask: Frontier for judgment, Mid-tier for delegation, Value-tier for autonomous background. Baseline is "everything on the frontier tier" — the most expensive but simplest architecture. Sextant's recommendation is the mix that preserves output quality at a fraction of the all-frontier cost.

Worked: Helios "Monthly financial close" — mix = $558/cycle (frontier $320 + 2× mid-tier $96+$128 + value $14); all-frontier = $1,880/cycle; per-cycle savings = $1,322 (70% lower); annual = $1,322 × 12 = $15.9K.

5.6 · Projected Impact panel · headline aggregation

Say it"This rolls the five computations into the four numbers your CEO repeats in the boardroom — where you land, what it costs, your single biggest lever, and how fast it pays back."

The "Adopting N tracks" card on the Command Center aggregates the first five computations into the four KPIs a CEO will quote back: composite readiness, total investment, top-leverage pillar, and projected payback. Each is a pure roll-up — no new math, just composition of 5.1–5.5.

Worked: Helios all 6 tracks → composite 3.15 (Δ +1.71), top lift = Value (1.3 → 3.6 = +2.3), 10,096 hrs reclaimed = 4.9 FTE-equivalents, $2.1M annual recurring savings, full-program payback ≈ 219 days (vs. the 49-day Drift-payback tile, which is the first track alone). Investment total is scoped per proposal and not shown here.

Limitations · when not to trust these numbers

Demo data is industry-typical, not company-specific. The Sextant Discover phase calibrates these formulas to a client's actual workflow telemetry (cycles, hours, rates, tool spend) within the 28-day engagement — at which point every value above reflects the client's reality, not a default. Until then: treat demo numbers as the shape of the answer, not the answer itself. Industry weights, decay rate (5%/quarter), commitment factor (0.65), and classification thresholds are documented in the methodology and reviewable on request.

Agent-readiness ladder · cross-cutting

Three tiers; clients walk up one rung at a time. Roadmap rule: never recommend a tier the client is not adjacent to.

Discover → Empower → Build

Sextant is the entry point of the delivery model, not a standalone product. Discover determines where AI should create value; Empower and Build turn that agenda into realized capability.

Discover, Empower, and Build are sequenced but not one-and-done — findings from a completed cycle routinely reopen the Discover agenda as priorities, capabilities, and constraints evolve.

External validation · public research

Every diagnostic figure surfaced to a CEO is anchored to public research the buyer can independently verify.