Authority at execution
Define who has authority over a workflow, data use, system action, or escalation point.
Pilot discussions
Pilot discussions for applied AI governance
OBEXGATE is working with early pilot partners to test how governance rules, authority boundaries, and audit controls can operate where AI and data actions are actually executed.
OBEXGATE is opening selected pilot discussions with organisations that need AI governance to move beyond policies, PDFs, and static checklists.
The focus is applied governance: how authority, consent, documentation, and escalation controls operate at the point where AI systems, agents, data workflows, or automated decisions are executed.
Why pilot with OBEXGATE?
Many organisations already have AI policies. The harder question is whether those rules can be enforced when systems are acting, routing, classifying, approving, denying, transcribing, summarising, escalating, or sharing data.
Operational governance controls
Test when an AI or data action should be allowed, warned, held, blocked, or stopped.
Audit-ready evidence
Produce structured records showing what rule applied, what action was requested, what control was triggered, and what human review occurred.
Real-world workflow testing
Use synthetic or controlled scenarios to test governance behaviour before live operational exposure.
Applied learning
Help teams understand how governance works in practice, not only how it is described in policy.
Current pilot traction
OBEXGATE has secured an initial pilot partner in New Zealand focused on Indigenous data governance and applied AI governance.
The pilot examines how authority, consent, provenance, AI transcription, synthetic clinical data, and secondary research use can be governed through operational controls rather than policy language alone.
Additional company agreements are awaiting signature before the partner can be named publicly.
New Zealand pilot focus
The New Zealand pilot is focused on a clinical patient use case using synthetic data. The scenario examines how information captured in a clinical context may later move through AI transcription, data handling, research preparation, secondary analysis, or external use.
OBEXGATE is being used to test how governance boundaries can travel with the workflow and trigger the appropriate control when a proposed action exceeds the permitted authority.
Clinical patient use case
A controlled clinical scenario for testing governance at the execution layer.
Synthetic data
Synthetic data enables governance testing without exposing private or live sensitive data.
AI transcription
AI transcription governance can be tested before operational use creates avoidable exposure.
Consent and authority boundaries
Controls can evaluate provenance, consent, authority, escalation, and secondary research use.
University pilot pathway
OBEXGATE also has a US university pilot pathway in discussion, focused on applied governance education for students.
The goal is to help students understand how AI governance operates in practice: how decisions are classified, what authority applies, when documentation is needed, when escalation is warranted, and how governance controls can be tested before systems are deployed.
Applied governance for students
A university pilot can expose students to practical governance problems using structured scenarios, synthetic data, and operational decision controls. This creates a bridge between AI policy education and the real-world systems organisations are now deploying.
Pilot awaiting signed agreements
The pathway is in discussion and will only be described publicly in more detail after appropriate agreements are signed.
What an OBEXGATE pilot can test
AI and data workflow governance
Test how rules apply when data moves across systems, teams, models, vendors, or use contexts.
Authority boundaries
Define who can approve, deny, escalate, or override a proposed AI or data action.
Consent and secondary use
Examine whether a proposed use remains within the original permitted purpose or requires additional review.
AI transcription and summarisation
Test governance controls around clinical, educational, administrative, or operational transcription workflows.
Synthetic scenario testing
Use synthetic data and controlled scenarios to validate governance logic before production deployment.
Audit and evidence trails
Produce structured evidence showing what happened, what rule applied, and what decision path was followed.
Student or workforce education
Translate AI governance from abstract policy into repeatable operational judgement.
From policy to control
OBEXGATE is designed around operational control states. A pilot can test how governance rules behave when a system action is proposed.
Within the approved boundary
The action is within the approved boundary and can proceed.
Proceed with governance notice
The action can proceed, but the user must be notified of governance risk or required documentation.
Review before proceeding
The action requires review, acknowledgement, or additional context before proceeding.
Outside the permitted boundary
The action is outside the permitted boundary and should not proceed.
Hard stop required
The action requires a hard stop because authority, consent, safety, or legal constraints are materially implicated.
Observer mode can be used where an organisation wants to identify governance issues without interrupting operations during early testing.
Who should discuss a pilot
OBEXGATE pilots are best suited for organisations that need to test AI governance in real workflows, not only write governance language.
Applied AI governance education
Student training, research governance, and practical scenario testing.
Healthcare and life sciences
Clinical workflows, transcription, prior authorisation, patient data handling, secondary research use, and audit controls.
Indigenous data governance organisations
Authority, provenance, consent, local control, and secondary-use governance.
Enterprise AI teams
Agentic workflows, internal AI tools, shadow AI, AI-assisted decisions, and governance escalation.
Data infrastructure and platform companies
Governance-aware data movement, evidence trails, authority controls, and operational oversight.
Public sector and civic institutions
AI adoption, service delivery, accountability, escalation, and public-interest governance.
What a pilot is not
A pilot is not a legal certification, a compliance guarantee, or a substitute for legal advice. It is a structured way to test how governance requirements, authority boundaries, and operational controls behave in realistic AI and data workflows.
Not a legal certification
A pilot does not certify legal or regulatory compliance.
Not a compliance guarantee
A pilot does not guarantee a regulatory outcome.
Not a replacement for legal counsel
Legal interpretation remains with qualified legal advisers.
Not a production deployment unless separately agreed
Production use, public naming, and case-study publication require separate agreement.
Not dependent on private live data
Pilots can use synthetic data or controlled scenarios where appropriate.
How a pilot discussion works
Scope
Define the workflow, system action, data movement, or decision pathway to be tested.
Classify
Identify the relevant governance issues: authority, consent, documentation, escalation, safety, auditability, or secondary use.
Simulate
Use synthetic data or controlled scenarios to test governance behaviour.
Evaluate
Review where controls triggered, where ambiguity appeared, and where escalation was needed.
Decide
Determine whether the pilot should move into a deeper implementation, educational programme, research collaboration, or strategic partnership.
Pilot governance model
Workflow context
A local clinical encounter or controlled workflow is the starting point for the governance test.
Authority and consent checks
OBEXGATE evaluates whether AI transcription, synthetic data handling, or secondary use remains within the approved boundary.
Provenance record and audit trail
The pilot tests whether governance decisions can produce structured records of action, control, review, and outcome.
Allow, Warn, Hold, Block, Stop
Proposed actions can be evaluated through operational control states rather than policy text alone.
Discuss a pilot with OBEXGATE
We are currently prioritising pilot discussions where governance risk is operational, cross-system, or difficult to resolve through policy alone.
Suitable pilots may involve AI transcription, agentic workflows, synthetic data, student training, healthcare governance, Indigenous data governance, research use, authority boundaries, or enterprise AI controls.