Evidence-linked reliance assurance
RelianceCheck gives consequential work a private, evidence-linked second look. It surfaces what needs attention, shows why, and points to the next corrective step—whether the work was made by a person, an AI, or both.
Public explanation active · document processing remains private
Illustrative interface
Claim
“The proposal guarantees a 30% reduction in year one.”
Evidence gap
No baseline, calculation method or supporting source is supplied.
Reliance effect
Do not rely on this claim yet. The conclusion outruns the evidence.
Corrective next step
Add the baseline, method, accountable owner and source—or qualify the claim.
01
Two moments, one question
Work coming in
Work going out
02
From finding to action
Start with the sentence, number or conclusion that needs attention.
Show what supports it, contradicts it, or is still missing.
Explain what the issue changes and why it matters to the decision.
Give a precise prompt for correction, clarification or qualification.
03
What RelianceCheck challenges
Conclusions that move further than the evidence beneath them.
Statements, assumptions or calculations that do not fit together.
Evidence whose date, provenance or relevance needs another look.
Unsaid limits, undefined terms and omissions that can change the decision.
The honest boundary
RelianceCheck is source-neutral: human-made, AI-made or mixed. It is not an AI detector, does not certify truth, and does not replace professional judgement. Its purpose is to make reliance risks and their evidence easier to inspect.
Reactivation status · 14 October 2026
This site explains the service and accepts no documents, accounts or payments. Assurance execution remains separately gated while the controlled pilot is re-provisioned. No visitor content is processed on this public property.
