A1AYN Intelligence
Methodology
How A1AYN structures analysis, labels confidence, and grounds outputs in traceable sources. This page is the summary. The full epistemic contract, six objects from entity to forecast with the evidence rules for each, lives on the data methodology page. Scoring rubrics and data source citations are documented in each domain's own methodology page.
Principles
Analytical Philosophy
- Evidence first
- Claims are separated from indicators. Every output distinguishes what a source directly states from what has been inferred from consistent signals.
- Traceable sources
- All analytical inputs are cited. Where a source is not publicly accessible, this is noted explicitly.
- Explicit assumptions
- Assumptions are stated, not buried. Where data is incomplete, the gap is named rather than papered over.
- Reproducible outputs
- Versioned data and documented processing steps mean outputs can be re-run and audited.
- Confidence labeling
- Every output is written to make uncertainty visible. See the Confidence labeling section below.
Architecture
The Six Objects
Every data surface on the site is built from six objects. Each is defined, with its evidence rules, on the data methodology page; the links go to live pages where each object is visible.
- Entity
- A country, company, facility, product, water body or bloc in the canonical registry. Everything else attaches to entities.
- Claim
- A statement typed as descriptive, causal or predictive, and labelled measured, inferred or projected. Causal claims carry an identification class.
- Observation
- A sourced data point with its vintage. The raw material claims are tested against.
- Dependency
- A typed edge between entities: what requires what, with mechanism, sign, lag and evidence. AND-groups capture gates that need every input at once.
- Model
- A transfer function with stated assumptions and backtests, like the Finnish merit-order price model.
- Forecast
- A claim with a falsifier and a resolution date, logged before the outcome and scored publicly when it resolves.
- Full contract
- Definitions, evidence rules and the identification classes for causal edges, with links to every live implementation.
Scope
What We Don't Do
Understanding the boundaries of this service is as important as understanding its capabilities.
- No automated alerts
- Outputs are analytical products, not live monitoring systems. Nothing here triggers automatically.
- No real-time feeds
- Data is updated on documented schedules. Nothing on this site claims to reflect conditions at the moment of viewing.
- No legal advice
- Analyses are indicative intelligence for sourcing and procurement decisions. They do not constitute legal, regulatory, or financial advice.
- No black-box scoring
- Every score on this site has a documented methodology page explaining exactly how it was derived and from which sources.
Interpretation
Confidence Labeling
Outputs are written to make uncertainty explicit. Every claim on a data page carries one of three epistemic tiers; hypotheses and causal claims carry additional obligations.
- Measured
- Directly supported by the cited source data. The source makes the claim; we are reporting it.
- Inferred
- Derived from consistent signals across multiple sources but not a direct statement of any single source. Reasoning is shown.
- Projected
- A model output or scenario anchored to measured capacity or observed behaviour, with assumptions stated. Never presented as a measurement.
- Hypothesis
- A plausible interpretation that requires verification. To graduate, it must be registered as a forecast with a falsifier and a resolution date in the Forecast Lab.
- Causal claims
- Additionally carry an identification class: statutory, forward test, natural experiment, or mechanism. Correlation alone is inadmissible.
Scoring Rubrics
Domain Methodologies
Each analytical domain maintains its own methodology page documenting the scoring framework, all data sources with citations and vintages, band definitions, and update cadence. This is where scoring rubrics live.
- Data & evidence architecture
- The six objects, the epistemic tiers, identification classes for causal edges, and the reference implementation. The contract behind every data page.
- EU Compliance & Market Access
- Country risk scoring across nine dimensions. Sources: TI CPI 2024, US DOL ILAB 2024, EUDR benchmarking May 2025, EU IUU carding records, LRQA EiQ 2024, EcoVadis 2025.
- More domains
- Methodology pages for additional domains will be added as each domain's scoring framework is formalised.
Caveats
Limitations
- Data delays
- Public data sources may be delayed, revised, or incomplete. Update cadence for each source is documented in domain methodology pages.
- Rules of origin
- Rules of origin and manufactured-in claims require case-by-case verification and are not determined by country-level scores.
- Indicative only
- Analyses are indicative intelligence. They do not substitute for legal review, supplier audits, or formal due diligence.
- Sector variation
- Country-level scores are starting points. Risk within a country varies significantly by sector, product category, and supply chain tier.