DMGT Economic Cognition Ledger

Metering intelligence across Training · Inference · Personalization · Platform · Subscription — weighted by a live Trust Index. Adjust the model and every number recomputes instantly.
LIVE · ledger-reconciled
PEL FEED

Portfolio Snapshot

recomputes with the what-if engine below

Economic Strategy Matrix

daily $ value per data class × strategy — click a row to drill in
Data class All
Strategy columns Training Inference·RAG Personalization Platform Subscription

Daily value by strategy

Where the money concentrates across the portfolio

Value contribution by data class

Daily total share

Trust Index (GSTI)

provenance · freshness · compliance · exclusivity · verification → Trust Multiplier θ

GSTI dimension heatmap

0–10 per dimension · composite drives θ on every ledger line
High (≥8.5) Mid (6–8.5) Low (<6)

Trust Multiplier θ by class

θ = 0.6 + (GSTI / 10) — applied to every strategy

Telemetry Findings

what Canary + Sentinel detect at the inference layer
robots.txt efficacy
1 of 18
chatbots actually blocked
Inference-layer detection
94%
of consumption events attributed
Dual-cosigned lines
764
of 1,493 PEL rows
SoR-reconciled value
$7.14M
settlement of record

Detection coverage by data class

Canary coverage · spoof-resistant · dual-cosigned · compliance α

Key findings

The evidence base for every Trust Index score

Agent-Workflow Fit

which class × strategy works best for agents — Agent Fit Score 0–10

Agent fit heatmap

Route agent workflows to the green cells first; starve Tier-1 commodity
Strong fit (≥8.5) Moderate (6–8.5) Weak (<6)

Predictive Economic Ledger (PEL)

30-day projection the inference agent emits — SoR-reconciled, dual-cosigned

Daily metered value · 30-day projection

Stacked by strategy · responds live to the what-if model

Intelligence Economics — Core Measurements

metered on a live cross-vendor AI workforce window · 1,493 decisions · two frontier runtimes on every one
Price per Intelligence Unit
$12.67
fully loaded · vs $50.67 human-only baseline
Baseline multiple
4x
lower cost per unit than human-only
Total intelligence cost
$18,921
model + coordination, full window
Raw model cost / unit
$0.0066
the intelligence-cost floor
Intelligence units
1,493
250 qualified · 401 disqualified
Coordination tax
32%
of output value — the autonomy lever
Throughput
12.3/day
sustained across the window
Realized outcomes
$7.14M
SoR-reconciled · Return on Intelligence next

Where the intelligence dollar goes

Model spend is the floor; coordination is the cost — and the lever autonomy graduation pulls

Per-unit economics

Human-only baseline vs AI fully loaded, per qualified decision

Intelligence units per week

Real metered volume across the 31-week window

Raw model cost per unit — weekly

The floor holds under one cent while volume varies

What-If Engine

adjust rate card, consumption scalars, reach & buyers — everything above recomputes
Editable model inputs — the blue logic of the ledger