How it works
From one verified company identifier to a decision-ready report.
MadeSure AI turns scattered public information into a structured supplier investigation while preserving where every claim came from and how strongly it was verified.
MadeSure AI investigation flow
- Unified Social Credit Code (USCC)The unique 18-character identifier shown on the supplier’s Chinese business licenceRegistry verification
- 北京示例有限公司AI research
Global-facing sources
- Global B2B Platforms
- Global Social Media
China-local sources
- Chinese Social Media
- Public Procurement Records
- Chinese B2B Platforms
- Customs & Trade Records
- Official Company Registries
- Legal & Enforcement Records
- Recruitment Platforms
AI synthesis- AI Supplier Due Diligence Report
AI-powered supplier intelligence, shaped by professional sourcing expertise, with source-linked findings on identity, manufacturing, trade activity, and risk.
View Sample Report
One code. One legal entity. Clearer next steps.
WHAT IT IS
An 18-character registration identifier
The Unified Social Credit Code identifies a registered Chinese legal entity. It is not a credit score and does not, by itself, prove that a supplier is reliable.
WHERE TO GET IT
Ask for the Chinese business licence
The code is usually printed on the supplier’s Chinese business licence. Ask the supplier to send the licence or provide the complete 18-character code.
WHY IT MATTERS
Match the entity before you research it
Use the USCC to compare the registered company with the quotation, contract, invoice, payment beneficiary, and other supplier documents.
Four stages, one evidence trail.
- 01
Verify the legal entity
MadeSure AI sends the submitted 18-character USCC to the company registry and starts deeper research only after the same code resolves to one mainland-Chinese legal entity.
- 02
Collect China-native and global evidence
We gather source-linked evidence across Global B2B Platforms and Global Social Media, then China-local sources including Chinese Social Media, Public Procurement Records, Chinese B2B Platforms, Customs & Trade Records, Official Company Registries, Legal & Enforcement Records, and Recruitment Platforms.
- 03
Use AI to reconcile claims and conflicts
AI compares supplier claims with independent evidence, connects related entities, identifies contradictions and patterns across sources, and keeps uncertainty visible instead of turning a missing signal into a negative fact.
- 04
Turn the evidence into a decision-ready report
AI synthesizes the collected evidence into a structured report covering identity, manufacturing signals, trade activity, risks, limitations, and recommended next questions, while keeping every important finding traceable to its source.
What the report does not do: it does not guarantee that a supplier is safe or replace product inspections, legal review, or a physical factory audit.
Nine source families, cross-validated into one conclusion.
MadeSure AI collects relevant evidence from these nine source families and analyzes each item against the resolved legal entity. It then cross-validates names, dates, locations, products, relationships, and operating claims across sources, reconciles conflicts, and produces the most reliable conclusion the available evidence supports—while keeping limitations and unresolved questions visible.
- 01Collect
Search relevant source families using the resolved legal name, identifiers, aliases, domains, and account handles.
- 02Analyze
Extract entities, dates, products, locations, activity, risk events, source URLs, confidence, and access limits.
- 03Cross-validate
Compare supplier claims with independent records and test identity, relationships, timelines, and operating signals for consistency.
- 04Conclude
Reconcile conflicts, weigh source strength, and produce the most reliable conclusion the available evidence supports.
Public storefronts, product focus, platform history, MOQ, badges, ratings, reviews, and response metrics.
Tests product focus, selling history, and market-facing activity; storefront data remains supplier- or platform-displayed.
Public company profiles and activity across LinkedIn, Instagram, Facebook, YouTube, TikTok, X, Reddit, and similar platforms.
Compares export-facing identity, content, and recency while separating official, employee, partner, and third-party accounts.
Public WeChat and China-native posts showing products, factory activity, exhibitions, hiring, projects, and recent updates.
Adds local, time-stamped signals and connects Chinese names and accounts; ownership and post authenticity are checked separately.
Public procurement notices, tender awards, project records, buyer names, products, locations, dates, and contract signals.
Shows China-local delivery history and product or project relevance where public notices are available; absence is not evidence of no activity.
Public storefronts, product focus, platform history, MOQ, badges, ratings, reviews, and response metrics.
Tests product focus, selling history, and market-facing activity; storefront data remains supplier- or platform-displayed.
Shipment dates, shippers, buyers, products or HS codes, routes, quantities, and destination markets.
Tests claimed export experience, products, buyers, and destinations within provider coverage—not total exports or future performance.
Legal name, registration ID, establishment date, address, business scope, representative, and operating status.
Anchors every later finding to the correct company and exposes mismatches in names, identifiers, addresses, or transaction roles.
Business anomalies, penalties, serious violations, court matters, enforcement, restrictions, bankruptcy, and equity freezes.
Places records in context by entity, role, date, and status; incomplete coverage or no record is never treated as no risk.
Public hiring for export sales, production, quality, engineering, and operations, including location and timing.
Role mix, location, and timing can support manufacturing, quality, engineering, or export functions, but not headcount or capacity.
Coverage is reported, not assumed. A source may be found, access-limited, unavailable, not found, or outside scope. “Not found” never becomes “does not exist.”