PUBLIC AI TRANSPARENCY PROFILE
Meridian
Understand the assistant—and the practices behind it.
AgenticertAI TRANSPARENCYView disclosure A BASIC FLOW MAP, INSIDE THE DISCLOSURE
One request.
Visible responsibilities.
This fictional assistant retrieves company information, prepares a draft through a tool, and waits for a person to approve it. Select a step to inspect its related disclosure.
DISCLOSED PRACTICE / 01
Explain the AI feature
The organization describes the task this feature supports, the people it serves, and the limits of its use.
- Information involved
- The request and account context enter the assistant.
- Responsibility
- Application owner
- Basis
- Company-declared example · not independently assessed
These statements and controls are illustrative company declarations. No operational behavior, supporting evidence, or approval mechanism has been independently assessed.
- Human oversight Declared
- A designated person reviews the draft before sending.
- Data access Declared
- The assistant uses selected company records; the access configuration is not supplied.
- Testing Not supplied
- No evaluation results or review record have been attached.
- Incident response Not supplied
- An incident contact, escalation policy, and response timelines remain unknown.
THE PRACTICES BEHIND THE WORKFLOW
More than the data journey.
Data handling
Information categories, purposes, providers, locations, and retention—with unknown details made explicit.
Security
Disclosed permissions, access boundaries, and safeguards, linked to supporting records where available.
Human oversight
Who approves actions, who can intervene, and who owns the decisions when a workflow changes.
Testing
What was evaluated, when it was reviewed, and the limits of what those results establish.
Incident response
How problems are reported, escalated, investigated, and communicated to affected people.
Accountability
A defined disclosure scope, responsible owner, and review date so readers can judge the context.
THE DISCLOSURE, IN PLAIN LANGUAGE
What this example covers.
These are sample statements for a fictional product. The workflow illustrates where each practice applies. Provider identities, exact processing regions, retention periods, evaluation results, and incident contacts would need to be supplied and supported in a live disclosure.
What information is involved?
Questions, selected document excerpts, account identifiers, and operational events. Different services receive different categories; inspect the connection details to see the intended transfer.
Which services are involved?
The illustrative workflow includes the Meridian application, access to approved company records, a connected drafting tool, and a designated human reviewer. A complete disclosure would identify each provider and its role.
Where is information processed or stored?
Exact processing and storage locations are not supplied in this illustrative workflow and remain unknown. A real profile should record them separately from company headquarters and identify the source behind each statement.
What safeguards are described?
This example declares access to approved records and a human approval checkpoint before sending. No test report, access-control record, or independent assessment is provided; these are declarations rather than verified controls.
What is outside the scope?
This example excludes actual message delivery, billing, staff systems, and other product features. It does not assert that the fictional application is secure, compliant, independently audited, or suitable for a particular use.
SUPPORTING CONTEXT
The basis of this profile.
No real evidence has been assessed. The materials below illustrate what a live disclosure could reference.
Architecture declaration
Services, intended dependencies, and data paths.
Data-handling record
Categories, purposes, locations, and known gaps.
Independent corroboration
No independent assessment is provided for this fictional profile.