About FactGuard AI
Supply chain experience behind the algorithm.
Founder-led AI development. Human judgment, retained.
FactGuard AI is developing an AI-assisted cross-checking algorithm for authorised supplier documents, designed around procurement experience, human-in-the-loop testing and traceable evidence.
The intended fact packs support preparation for negotiation, supply chain due diligence, risk mitigation and informed decisions. The algorithm is under development; its performance is not presented as independently validated.
Why FactGuard exists
A supplier claim is not yet decision-ready evidence.
Supplier decisions depend on the information available—not only the information a buyer would like to have.
In a 2022 CIPS–Deloitte survey of 468 procurement and supply-chain professionals, 71% reported limited or no visibility beyond tier 2.
Network visibility and document reliability are different challenges. FactGuard’s initial focus is on the documents a buyer is authorised to review: what they support, where they differ, and what still needs clarification.
FactGuard was founded to bring an evidence-focused approach to supplier information. The aim is to organise documented support, interpretation and unresolved questions into traceable fact packs for human decisions.
Source note: CIPS–Deloitte, 2022; 468 survey respondents. The finding measures reported supply-chain visibility, not document accuracy or FactGuard performance.
Meet Nelson
Nelson
Founder & Algorithm Designer
Former Fortune 500 Supply Chain Executive
Nelson brings an engineering background, an MBA and experience in international procurement and supply-chain leadership. His work has included global category strategy, supplier selection and assessment, commercial negotiation, cost governance, supplier development and procurement digitalisation.
That technical and commercial perspective informs FactGuard AI’s approach: the questions an algorithm should help cross-check, the evidence that needs context, and the points that must remain subject to human judgment.
Nelson is developing FactGuard’s AI-assisted cross-checking algorithm to support the review of supplier-supplied documents, with human-in-the-loop testing and review as part of its design. The aim is to support professional decision-making—not replace it.
How the planned method works
AI-assisted document cross-checking
The algorithm is being developed to support comparisons across authorised supplier documents and identify points that need further review.
Human-in-the-loop testing and review
The planned review process examines outputs, source support, context and unresolved questions. Algorithmic output is not intended to become a final finding simply because it has been generated.
A traceable fact pack
The intended output separates documented support, interpretation and open questions in a form that people can review, challenge and use when preparing decisions.
What guides the development
Evidence before assertion.
Keep a claim connected to its supporting material.
Human accountability.
Technology assists the review; professional judgment remains responsible for decisions.
Transparent limitations.
Missing or conflicting evidence should remain visible, not be converted into unwarranted certainty.
The intended fact packs do not replace an audit, certification, legal review or supplier qualification decision. Any pilot scope, materials and human-review arrangements are agreed separately.
Discuss a focused pilot
Tell us what decision you are preparing for. We can discuss fit, development-stage limitations and a possible pilot scope before an engagement is agreed.
Request a PilotContact
Contact
FactGuard AI is operated by Expert Connected Hong Kong Limited.
- Business correspondence address
- 7/F, Kowloon Building, 555 Nathan Road, Kowloon, Hong Kong
Enquiry form
Share a brief, non-sensitive question. Please do not include confidential documents or files.