Skip to content

Solutions · AOKAI

atestia

A product of AOKAI

AI, declared. Work, trusted.

Atestia helps universities define and publish their AI use policy, and lets each professor set what level of AI is allowed in each assignment. Students declare how they used AI, directly from their assistant or through a form, and the institution gets data and training to promote honest, responsible use.

Talk with usSee the modules

The problem

Policy on paper. Rules improvised. Trust left to detectors.

Most institutions have an AI policy on paper, or none at all. Professors invent rules per course, students guess what is allowed, and integrity disputes lean on AI detectors that are unreliable. Leadership has no data on how AI is actually used.

  • Institutional policy that few people have read or accepted
  • Course rules that vary by professor and stay invisible to students
  • Detectors that raise false accusations and miss honest use
  • No aggregated evidence for deans, rectors, or procurement

Declare, don't detect.

How it connects

From policy to disclosure — then back again

The policy sets the limits. Professors apply them per assignment. Students disclose against them. What you learn feeds training and the next policy version.

  1. 01

    Policy

    Draft, publish, and track acceptance.

  2. 02

    Course rules

    Assign an AI level to every assignment.

  3. 03

    Disclosure

    Students declare how they used AI.

  4. 04

    Insights

    See patterns without exposing individuals by default.

  5. 05

    Training

    Teach responsible use where the data points.

Insights and training shape the next policy.

Five modules

Everything the campus needs to declare AI use

Policy and acceptance

Leadership drafts and versions the institutional AI policy. Every audience accepts the current version — with a plain-language summary, optional comprehension check, and an audit trail.

  • Templates aligned with international guidance
  • Versioning, review workflow, and multilingual text
  • Acceptance by faculty, program, or the whole community

Course-level AI rules

Professors set a default AI level for the course and overrides per assignment. Students see a plain card: what is allowed, and how to disclose.

  • Assignment codes and QR for quick access
  • Rules cannot exceed the institutional maximum
  • Works standalone — a professor can start without a contract

Student disclosure

Students declare AI use through an MCP connector, a copy-paste summary, or a manual form. Nothing reaches the professor until the student reviews and submits.

  • Same disclosure record from every route
  • Flags when declared use exceeds the assignment level
  • Opens a conversation — never an automatic penalty

Usage insights

Aggregated evidence for professors and leadership: who disclosed, what kinds of help, tools in use, and where rules may be unclear.

  • Anonymized by default above a minimum group size
  • Course views for faculty; institution views for leadership
  • Export to CSV and PDF for accreditation

AI literacy training

Short practical courses for students, professors, leadership, and staff — connected to policy acceptance and to insights that need a refresher.

  • Micro-lessons with quizzes and certificates
  • Tracks for basics, responsible use, teaching, and governance
  • Spanish, English, and Portuguese at launch

AI use scale

Clear levels for every assignment

The default scale is editable in the institutional policy. Levels 1 to 4 always require disclosure. A professor cannot set a level above the maximum allowed for that context.

RestrictOpen
  1. 0

    No AI

    No AI tools at any stage.

    No AI allowed
  2. 1

    AI for planning

    Brainstorming, outlining, understanding the task — no AI-generated content in the submission.

    Disclosure required
  3. 2

    AI for feedback

    AI may review and suggest improvements to the student's own work.

    Disclosure required
  4. 3

    AI-assisted work

    AI may generate parts of the work, which the student edits and evaluates.

    Disclosure required
  5. 4

    Full AI

    AI used freely as a collaborator; critical evaluation is part of the grade.

    Disclosure required

Levels 1–4 always require disclosure.

Three ways to disclose

Assisted, student-approved disclosure

Every route produces the same disclosure record. The student always reviews the summary before it reaches the professor.

MCP connector

Students connect Atestia to their AI assistant in an institutional workspace. The assistant receives the assignment rules, logs meaningful help, and drafts a disclosure summary.

Copy-paste summary

Paste a prompt into any assistant, get a structured summary of the help in that chat, and paste it into the disclosure form. Works with every assistant, including free plans.

Manual form

The universal fallback: tools used, purpose, extent, and optional chat export.

What we state openly

  • The connector only records what the assistant reports; it does not observe the conversation.
  • It only works in assistants that support MCP connectors with write actions.
  • A student can use AI outside a session — so this is assisted disclosure, not monitoring and not proof.

The connector works where the workspace supports write actions; the summary and the form cover everyone else.

Who it serves

Built for the people who teach, study, and lead

Students

Your work, your word.

Know what is allowed, disclose AI use easily, and avoid unfair accusations.

Professors

Start alone.

Set clear AI rules per course and assignment, review disclosures, and begin without an institutional contract.

Leadership

Trust you can show.

A published policy, proof of acceptance, and institution-wide usage data.

Quality and integrity

Fair process.

An audit trail and individual records only when a formal case requires them — with access logged.

IT and procurement

Tools in view.

Visibility on which AI tools are in use across the institution.

Principles

How Atestia makes decisions

Transparency over surveillance

Students always see and approve what is recorded about them.

Professor-first adoption

A professor can start alone, without an institutional contract.

Policy is the source of truth

Course rules can never exceed the institutional policy.

Privacy by default

Insights are aggregated and anonymized unless a specific case requires otherwise.

Tool-agnostic

Works with any AI assistant, with richer support for MCP-capable ones.

A conversation

Talk with us about AI governance on your campus

If you lead academics, quality, or digital transformation and want to replace detect-and-punish with declare-and-understand, leave your details.

Trust you can show.

A product of AOKAI · Codexsw Inc

Required fields. We will reply as soon as we can.