Methodology

How we source, verify, and maintain regulatory intelligence across 195+ jurisdictions.

1 How Facts Are Sourced

Every regulatory fact starts with primary research. Our autonomous pipeline queries multiple sources to build each jurisdiction's profile:

  • Official regulator websites — Central banks, financial authorities, securities commissions, and AML agencies for each jurisdiction
  • Primary legislation — Published laws, regulations, and statutory instruments from government gazettes and legal databases
  • Perplexity web research — AI-assisted searches cross-referencing multiple sources to identify current regulatory positions, recent changes, and enforcement patterns
  • FATF mutual evaluations — Country assessments from the Financial Action Task Force and FATF-style regional bodies (FATF, APG, CFATF, MONEYVAL, EAG, ESAAMLG, GAFILAT, GIABA, MENAFATF)

Each fact is stored with its source URL, source authority, and the date it was collected. Facts without verifiable sources are flagged as "unverified" until confirmed.

2 Verification Process

Verification is the process of confirming that a fact's claim still holds true against its cited source. Our pipeline runs automated verification cycles:

Weekly URL Checks

Source URLs are checked for availability. Dead links trigger re-verification or status downgrade.

Source Confirmation

Perplexity re-checks each claim against the cited source to confirm accuracy. Claims that no longer match are flagged as "changed" or "stale".

Cross-Reference

Key facts are validated against multiple independent sources when available, increasing confidence scores.

Staleness Detection

Facts not re-verified within 90 days are automatically downgraded. Rapidly changing jurisdictions have shorter thresholds.

3 Confidence Scoring

Every fact carries a confidence score from 0.0 to 1.0. This score reflects our assessment of how likely the fact is to be accurate and current:

Score Status Meaning
0.8 - 1.0 Re-verified Confirmed by multiple verification cycles, source URL active, claim matches source
0.7 First verified Initially verified against primary source, awaiting re-verification cycle
0.5 Unverified Sourced from research but not yet independently verified
0.3 Stale Previously verified but source is no longer accessible or claim may have changed
0.0 - 0.2 Disputed / Revoked Contradicted by newer evidence or officially superseded

The green verification badge () appears only on facts with confidence 0.7 or higher.

4 Data Freshness

Regulatory data has a shelf life. We track freshness at multiple levels:

Green

Updated within 7 days — data is current

Yellow

Updated 7-30 days ago — data is likely current but may have changed

Red

Updated more than 30 days ago — data should be independently verified before relying on it

Each country page displays a freshness badge so you can immediately assess how current the profile is. The intelligence feed shows relative timestamps for at-a-glance recency.

5 Autonomous Pipeline

Web3 Compliance AI operates a fully autonomous research and verification pipeline:

  1. Research dispatch — New jurisdictions and topics are queued as research tasks routed to GPU workers via NATS JetStream
  2. Fact extraction — Research results are processed into structured fact records with source attribution
  3. Daily source monitoring — Automated checks verify source URL availability and detect changes
  4. Weekly validation — Facts are re-verified against their sources using Perplexity web research
  5. Continuous deployment — Verified facts are committed to the repository and deployed via Cloudflare Pages

The entire pipeline runs on self-hosted GPU hardware (~$30/mo electricity) with minimal API costs (~$3-8/mo for Perplexity web search). No human intervention is required for routine operations.

Questions about our methodology? Connect via MCP API for programmatic access to all facts with their confidence scores and sources.