About

Corrections and appeals policy

v1.0, published 2026-07-05. Every substantive change is logged in the quarterly notice-and-action report.

# Corrections and Appeals Policy

To file a correction, email info@trustgent.com with the subject line `Correction request: [provider or page]` and attach the evidence you want us to consider. Trustgent acknowledges every request within 5 business days and resolves or escalates within 20 business days. Substantive corrections are logged in the quarterly transparency report with the change, the reason, and the date. Methodology disputes go to the quarterly methodology review, not to per-listing correction.

This document sets out what can and cannot be corrected on Trustgent, how disputes are handled at each of three tiers, how corrections are recorded in public, and how EU Digital Services Act (DSA) Article 16 notices are processed on the expedited track. It sits alongside the methodology, the Trustgent Attestation Protocol, the editorial standards, and the summary of how we verify providers. Read together, the five documents describe the full audit chain from evidence to published listing to correction.

What can be corrected

Trustgent corrects factual errors on request. The categories below are the ones that qualify as a correction rather than a dispute over judgment.

Provider data errors. Wrong company name, wrong registered office, wrong contact email, wrong website URL, wrong founding year, wrong headcount band, wrong industry vertical, or wrong service description. If the underlying source (company registry filing, provider profile submission, public website) shows one value and the Trustgent listing shows a different value, the listing is corrected to match the source.

Misclassified verification level. If a provider believes the assigned tier (L0 through L5) does not reflect the evidence on file, the provider can request a re-audit under the appeal process described below. The evidence chain, not the request itself, decides the outcome.

Incorrect capability tagging. Trustgent tags providers by capability (for example, retrieval-augmented generation, evaluation harness design, computer-use agents, on-prem deployment). If a capability is tagged that a provider does not offer, or a capability the provider does offer is missing, the tag set is corrected against the underlying evidence.

Outdated evidence. If a case study, contract, or reference on which a verification decision rested has been superseded, withdrawn, or expired, the provider or the affected client can flag it. Depending on the tier, the listing is either re-audited or the tier is adjusted downward until fresh evidence is filed.

Typos and formatting. Small copy errors on any Trustgent editorial page are corrected without formality. These do not enter the public correction log because logging every typo produces noise that hides the substantive changes.

What is not correctable

The correction process is not a route to overturn evidence-backed editorial judgment. Two categories fall outside it.

Subjective evaluations backed by evidence. A provider disagreeing with an L2-versus-L3 assignment does not, on its own, override the evidence chain. The appeal path (below) is the correct venue for that disagreement, and it turns on whether new evidence exists, not on the provider's own assessment of its work. If the DD (due-diligence) agent has audited a case study to L2 because independent third-party corroboration is present but the client-authorized recorded proof required for L3 is not, the appeal cannot be resolved by argument. It is resolved by filing the missing proof.

Methodology disagreements. Requests such as "capability tagging should be less granular" or "the L3 evidence bar is too strict" are not per-listing corrections. They are proposals to change the methodology itself. The correct venue is the quarterly methodology review, published under /about/methodology with a change log. Submit these to the same email address with the subject line `Methodology proposal: [topic]` and they will be routed to the review queue rather than the correction queue.

Rank position. Trustgent's ranking function is plan-blind. It uses `verificationLevel`, `recordCount`, and `lastVerifiedAt`, enforced by a hardcoded allowlist in the ranking code. A provider cannot request a position change. A provider can request a re-audit of the verification level that feeds the rank, and if the audit succeeds the rank moves as a consequence. Paid features (branding, seats, analytics, bulk invites) never influence position, and no correction request can change that.

How to file a correction

Send an email to info@trustgent.com with the following in the message body.

1. The URL or provider slug the request concerns. 2. A one-sentence statement of the correction requested. 3. The evidence you want considered, attached or linked. For provider-data corrections this is usually a registry filing, a screenshot of the provider's own website, or a signed employee attestation. For verification appeals it is a case study, a contract, a client reference, or an auditable work output. 4. Your role: provider, client, buyer, third-party researcher, or member of the public. 5. Any deadline you are working to, for example a procurement decision.

Trustgent acknowledges every well-formed request within 5 business days. Acknowledgement includes a case number and the tier of process the request has been routed to. Requests are resolved or escalated to the appeal panel within 20 business days. If a request is malformed (no evidence, no URL, no clear ask), it is returned for completion and the SLA restarts when it is resubmitted.

Sensitive material (contracts, personal data, non-public deliverables) can be sent under a mutual non-disclosure agreement on request. Trustgent's default posture is to hold submitted evidence in confidence and to publish only the fact of the decision, not the underlying documents, unless the submitter authorises publication.

How disputes are resolved

Requests are routed to one of three tiers.

Tier A: editorial correction

Applies to factual errors, typos, formatting problems, wrong tags, and outdated links. Resolved by the editorial team without a panel. Turnaround is usually under 5 business days. If the correction is substantive (anything beyond a typo), a short note is added to the page indicating the date of the correction and its scope. The full record is written to the quarterly transparency report.

Tier B: verification appeal

Applies to disputed verification levels and disputed evidence outcomes. The DD agent re-audits the record with any new evidence filed in the appeal. The re-audit is reviewed by a panel of two Trustgent reviewers who did not perform the original audit. The provider (and, where relevant, the client whose evidence is under review) is notified in writing of the outcome and the reasoning. Turnaround is up to 20 business days from acknowledgement, with an extension of a further 20 business days permitted where reference calls have to be scheduled with third parties.

Appeals can go up or down. If new evidence supports a higher tier, the tier is raised. If the appeal exposes weakness in the original evidence, the tier is lowered. Appeals may also result in no change, in which case the reasoning is recorded and the record remains as originally verified.

Tier C: methodology dispute

Applies to disagreements with the verification standards themselves. Routed to the quarterly methodology review. The proposal, the reasoning, and the disposition are published under /about/methodology whether the proposal is adopted, deferred, or rejected. Contributors are credited unless they request anonymity.

Public log

The quarterly transparency report at /transparency publishes every substantive correction from the preceding quarter. Each entry lists:

The date the correction was made.
The page or provider affected.
The nature of the change (data correction, tier raised, tier lowered, capability re-tagged).
The reason.
Where relevant, whether the correction was self-initiated by Trustgent, requested by the provider, requested by a client, or requested by a third party.

Small copy corrections (typos, formatting) are not logged. Everything that changes what a buyer would rely on is logged. The public log is the accountability record for the plan-blind moat: if a paying customer's tier were ever adjusted in a way inconsistent with the evidence chain, it would show up here, and the log would show it inconsistent with the record. That is the point of publishing it.

DSA Article 16 notice-and-action

For content flagged as illegal under EU law, Trustgent operates an expedited process consistent with Article 16 of the EU Digital Services Act.

How to file. Send a notice to info@trustgent.com with the subject line `DSA notice: [URL]`. The notice must include: (a) a sufficiently substantiated explanation of the reasons the content is illegal, (b) a clear indication of the exact location of the content (URL), (c) the notifier's name and email address (unless the content concerns an offence for which anonymity is protected), and (d) a good-faith statement that the information in the notice is accurate and complete.

Timelines. Acknowledgement within 2 business days. Substantive response, including the decision taken and the reasoning, within 10 business days. Where the notice concerns a criminal offence involving a threat to life or safety, Trustgent will notify the relevant authorities without delay.

Decision. Content is either left in place, corrected, restricted (for example, delisted from indexes), or removed. The affected provider is notified with the reasoning and offered an internal complaint mechanism as required by Article 20. Statement-of-reasons submissions to the DSA Transparency Database are published where required.

Manifestly unfounded notices. Where a notifier submits notices that are manifestly unfounded, Trustgent reserves the right to suspend processing further notices from that source after a warning, consistent with Article 23.

Contact

Corrections queue: info@trustgent.com with the subject lines above.

Trustgent's verified badge is an editorial signal produced by the DD agent's evidence-chain audit. It is not a regulatory certification and does not constitute an EU AI Act conformity assessment.

How we verify, the six verification levels and the evidence chain that produces them.
The Trustgent Attestation Protocol, the versioned specification behind the marketplace.
Methodology, how the platform operates end to end.
Editorial standards, the voice and sourcing rules.
Transparency reports, the quarterly audit log.

Corrections policy v1.0, published 2026-07-05. Trustgent is operated by Rustenhoven Management B.V., KvK 92219500, Prinsengracht 234B, Amsterdam.