Brel Digital is an independent AI intelligence platform. Profiles, directories, and reports are built from publicly verifiable information. We do not invent founders, funding figures, product capabilities, or relationships. This page explains how selection, verification, review, and corrections work.
Verification process
Every company, founder, product, model, and technology association must be supported by at least one primary public source before it is marked ready for search indexing. Editors compare claims against company websites, official documentation, regulatory filings, peer-reviewed or vendor technical papers, and reputable news only when they cite primary materials.
- Entity identity: legal or brand name, official website, and stable public presence.
- What they build: products, models, or platforms described in first-party materials.
- Context fields: sector, geography, and technology tags only when evidence supports them.
- Relationships: founders, products, and models linked only when publicly evidenced.
Source hierarchy
- Primary: official company or product pages, documentation, model cards, press rooms, and regulatory notices.
- Secondary: reputable journalism that clearly attributes facts to primary sources.
- Tertiary: directories and aggregators — used for discovery only, never as sole verification.
When sources conflict, we prefer the most recent first-party statement and note uncertainty rather than invent a resolution.
Research workflow
Research begins with discovery (sector scans, documentation, model catalogs, and credible coverage). Candidates are drafted against a structured field checklist. Incomplete drafts remain published for readers when useful, but stay noindexed until the indexing readiness checklist is satisfied.
Entity lifecycle
- Draft / incomplete: visible on-site when published, search-noindex until ready.
- Verified / indexable: checklist complete and ready-for-indexing flag set.
- Updated: material changes refresh last-reviewed dates and sources.
- Corrected or deprecated: inaccurate claims fixed; obsolete entities redirected or unpublished when appropriate.
Editorial review
Human editors own what goes live. Automated assist may help structure research and surface public sources; it does not invent facts. “Why it matters” and similar framing are editorial assessments labeled as such — not paid placements.
See also our Editorial Policy for independence, disclosures, and audience standards.
Corrections process
Factual errors are corrected after source review. Stakeholders can request a correction via our Corrections Policy. Eligible fixes update the page and last-reviewed metadata when appropriate.
Update schedule
High-attention entities and sector hubs are reviewed on an ongoing basis. Broader catalogs (models, glossary, APIs) are refreshed in batch when upstream documentation changes. Last-reviewed timestamps appear on entity and insight templates when available.
Knowledge graph methodology
The Intelligence Graph connects companies, products, models, founders, industries, countries, technologies, reports, and insights. Edges exist only when supporting evidence is on file. We do not fabricate “related” links to densify the graph.
AI usage disclosure
Brel may use AI tools to accelerate drafting, categorization suggestions, and internal QA against public text. Final publication decisions, verification of facts, and readiness-for-indexing remain human responsibilities. AI output is never treated as a primary source.
Human review policy
No entity is marked ready for indexing without human review of the checklist fields and cited sources. Thin or speculative pages stay noindexed. Directory filters and combination explorer views are noindexed so search engines prioritize canonical hubs and profiles.
Contact
Questions about methodology: Contact Brel Digital or email contact@breldigital.com.