How we moderate

No report is published automatically. Every submission enters a moderation queue where a human reviewer checks three things: is the reporter identity-verified, does the evidence support the claim, and does the narrative comply with our content guidelines.

Reports that pass all three become published claims — the unit that affects trust scores. Reports that fail are returned to the reporter with an explanation. This is not algorithmic moderation. A person reads every submission.

Personal names, phone numbers, email addresses, physical addresses, and any personally identifiable information about individuals (not businesses) are automatically flagged and manually redacted before publication.

Reports that name individual employees by first and last name are edited to remove the name. The complaint stays. The personal identification does not.

Medical details, financial account numbers, and legal case identifiers are also redacted. The goal is to describe behavior, not expose private information.

We accept: receipts (physical or digital), invoices, order confirmations, email or text correspondence with the business, photos of delivered goods, screenshots of booking confirmations, contracts, and statements of work.

We do not accept: hearsay, secondhand accounts, social media posts as sole evidence, or reports with no supporting documentation. You must have had a direct transaction with the business.

All accepted evidence types are treated equally in terms of verification status. What matters for scoring is whether the claim is published, the severity level, and whether the business responds and resolves.

All uploaded evidence — receipts, invoices, screenshots, correspondence — is anchored through the EMILIA Protocol, which creates a cryptographic fingerprint and timestamp at the moment of submission. Once anchored, evidence cannot be altered by anyone — not the reporter, not the business, and not REKKN — without the change being detectable.

This means when a business disputes evidence, we can prove the evidence existed in its exact form at the time of submission. And when a reporter claims they submitted something, we can verify whether they actually did.

If a report is found to contain fabricated evidence, misrepresented facts, or is filed by someone who did not have a real transaction with the business, the report is removed and the reporter receives a warning.

Second offense: the reporter's account is suspended for 90 days. Third offense: permanent ban and revocation of the RK number.

Businesses can flag reports they believe are false. Flagged reports are re-reviewed. If the flag is substantiated, the report is removed. If the flag is not substantiated, the report stays and the flag is documented.

When a verified report is published, the business is notified by email. They have 7 calendar days to respond publicly through their dashboard or claim page.

The business response is published alongside the original report, unedited, as long as it does not contain personal attacks, threats, or doxxing. The response becomes a permanent part of the record.

If the business does not respond within 7 days, their silence is recorded. The report is marked “No response” and that status is visible to every future viewer.

Any business can appeal a report by submitting counter-evidence through their dashboard. Appeals are reviewed within 5 business days.

Valid grounds for appeal: the reporter was not a customer, the evidence is fabricated, the described issue was already resolved before the report was filed, or the report contains factual inaccuracies that can be disproved with documentation.

If an appeal is upheld, the report is removed and the reporter is notified. If an appeal is denied, the business receives a written explanation. Denied appeals can be escalated once to a senior reviewer.

If a single reporter files more than 3 reports against the same business within 90 days, the additional reports are held for enhanced review. Coordinated campaigns (multiple accounts filing similar reports in a short window) are flagged by pattern detection and reviewed manually.

If a business receives a sudden spike in reports that do not match historical patterns, a “review integrity hold” is placed and the reports are reviewed before publication.

The reporter does. When a business responds to a report, the reporter is asked: “Has this issue been resolved to your satisfaction?” If the reporter confirms, the report is marked “Resolved.” If the reporter says no, it stays “Unresolved.”

Businesses can also submit evidence of resolution (refund confirmation, replacement tracking, etc.). If the evidence is clear and the reporter does not respond within 14 days, the report is marked “Resolved by business — awaiting reporter confirmation.”

We do not sell advertising to businesses on the platform. We do not offer paid report removal. We do not allow businesses to pay for higher trust scores. We do not suppress reports based on legal threats unless compelled by a court order, in which case the suppression itself is noted on the record.

We do not make editorial judgments about who is “right.” We present the evidence, the response, and the outcome. The public record speaks for itself.

This page will evolve. As we encounter new types of disputes, edge cases, and adversarial behavior, we will update the rules and document every change publicly. The goal is not perfection on day one. The goal is transparent, improvable fairness.

If you believe a moderation decision was wrong, or if you have a question about how a specific report was handled, contact moderation@rekkn.ai.