AI decisions need an operating model.
Schools are being asked to govern fast-moving AI use while protecting human judgement, student learning, assessment validity and sensitive information.
Core Education Design helps leaders turn fragmented tools, uncertain rules and uneven staff capability into a coherent, school-specific system for responsible action.
Available as an independent engagement or integrated with Learning and Organisational Architecture as part of a wider school transformation.
Access has arrived faster than governance and capability.
AI is already influencing planning, communication, assessment, administration and student work. The risk is not simply inappropriate use. It is allowing disconnected decisions to become the school’s de facto strategy.
Rules without a coherent rationale
Generic positions rarely distinguish low-risk support from high-consequence decisions or explain where human judgement must remain decisive.
Assessment without sufficient process evidence
Tasks designed around final products alone may no longer provide valid evidence of what a student knows, understands or can do.
Tools adopted before data decisions
Staff may use convenient services before leaders have determined what information is appropriate to enter, which terms apply or who is accountable.
Capability that varies by individual
Schools become dependent on enthusiasts while other staff receive isolated demonstrations rather than shared principles, safe workflows and sustained support.
What Intelligence Architecture means
Intelligence Architecture is the coordinated design of AI use, human judgement, data handling, assessment integrity and staff capability. AI operates as an enabling layer inside the school’s existing learning, leadership and administrative systems. It is not treated as a stand-alone product or strategy.
Begin with educational purpose. Constrain technology accordingly.
The architecture is built around decisions the school must own, not whichever platform is currently attracting attention.
Human authority
People remain accountable for professional, educational and high-consequence decisions.
Purpose before tool
Every use begins with a defined problem, intended benefit and proportionate risk assessment.
Minimum necessary data
Workflows are designed to avoid entering personal or sensitive information unless expressly approved and appropriately governed.
Visible learning
Assessment design protects evidence of student thinking, process, application and independent capability.
Transferable capability
Staff learn principles and reusable methods rather than becoming dependent on one tool or consultant.
The service is the architecture, not a generic AI tool.
Applications are selected or configured only after purpose, risk, data and human oversight are understood. Representative work may include:
Leadership intelligence and direction
One-off presentations for a principal, executive team or whole staff on national and international developments, likely directions, institutional implications and practical preparation. This can extend into a contextualised roadmap and procedures through an ongoing partnership.
Consultancy and capabilityAssessment integrity redesign
Reviewing vulnerable assessment patterns and designing stronger evidence of process, judgement, application, reflection and appropriate AI disclosure. The aim is assessment that remains valid on its own terms, rather than dependence on detection tools whose reliability cannot be assured.
Consultancy and custom configurationStaff workflow enablement
School-specific structures for resource design, differentiation, feedback, quality assurance, reporting and communication, with defined review points and human approval.
Consultancy, training and licensed accessTargeted AI applications
Configured assistants that encode agreed school requirements, instructional priorities and workflow boundaries. These may use a client-approved third-party platform and remain subject to that platform’s available controls and terms. Where a Core Education Design application or licence forms part of the proposal, that commercial interest is declared in writing before it is recommended, and no engagement is conditional on adopting it.
Declared interest · custom configuration and third-party platformStudent AI capability
Age-appropriate guidance for verification, bias, attribution, disclosure, productive prompting, critical evaluation and knowing when independent work is required.
Consultancy and capabilityGovernance and implementation roadmap
A concise, sequenced position connecting approved uses, restricted uses, data expectations, assessment practices, communication, staff development and review responsibilities.
Consultancy and capability transferCustom GPTs and Gemini Gems are deployment options, not the architecture.
They may be considered when the school has approved the relevant OpenAI or Google environment, accepted the applicable contractual and data-handling position, and confirmed that the proposed workflow is compatible with school requirements. If those conditions are not satisfied, Core Education will recommend a different configuration, a lower-risk workflow or no AI deployment.
Where students are the intended users, minimum age eligibility, account type, parental consent and system-authority approval are determined by the school and its system before any deployment. Core Education Design makes no representation on a school’s behalf about a platform’s suitability for students.
Clear decisions about data, contracts and accountability.
No blanket claim about privacy, data sovereignty or Departmental compliance can responsibly cover every deployment. The position depends on the selected service, account type, configuration, contract, data entered and governing requirements.
| Component | What may process data | Contractual relationship | Required control |
|---|---|---|---|
| Consultancy | Information supplied to Core Education for the agreed advisory work. | The school or client contracts with Core Education Design under the consulting agreement. | Scope, access, permitted information, deliverables and retention expectations are agreed for the engagement. |
| Custom configuration | The approved platform used to build or run the configured workflow may process prompts, uploaded content and outputs. | The client contracts with the platform provider unless a written agreement expressly states otherwise. Core Education configures the approved environment. | Client approval of the platform, account type, proposed data, permissions, testing and human review process. |
| Licensed Core access | The delivery platform and any connected services identified for that application may process user inputs and outputs. | The school receives a defined school-level Core Education licence. This is a Core Education Design commercial product, declared as such before it is proposed. Third-party terms continue to govern the underlying platform. | The licence scope, authorised users, platform dependencies, acceptable inputs and support boundaries are documented. |
| Existing school platforms | Microsoft, Google, OpenAI, the LMS or another approved provider may process data according to the relevant service and configuration. | The school, system or Department contracts directly with that provider. Core Education is not a party to those terms unless expressly stated. | The client confirms approval, permissions, data classification, applicable provider terms and any internal review requirements. |
| Human decision-making | AI output may inform a draft, recommendation or analysis, but it does not assume professional accountability. | Responsibility remains with the authorised school decision-maker. | Named human oversight, verification and approval before any consequential use. |
Practical rule: no personal, sensitive, confidential or student-identifiable information should be entered into an AI service merely because the tool is accessible. The client must approve the proposed use and platform. The workflow must then operate within applicable law, policy, contractual terms and school requirements, with proportionate human oversight.
Vendor neutrality: Core Education Design receives no commission, referral fee or reseller margin from any AI or platform vendor, and has no financial incentive to recommend one provider over another. Where a Core Education Design application or licence is proposed, that interest is declared in writing beforehand.
What Core Education Design will not claim
- Automatic compliance. No platform or configuration is represented as compliant with any policy, standard or regulatory requirement merely because it has been recommended.
- Endorsement. No AI service is represented as endorsed by the NSW Department of Education or any other education authority.
- Sovereign hosting. No third-party platform is represented as sovereign-hosted.
- Continuous availability or accuracy. No platform is represented as continuously available, and AI output is never represented as accurate by default. Verification remains a human responsibility.
- Certification. Core Education Design may identify security and privacy risk but does not provide cybersecurity certification unless separately contracted.
Build school capability, not consultant dependence.
The goal is a school that can make better decisions as platforms change. That requires transfer of reasoning, structures and governance ownership, not an expanding collection of opaque tools.
Platform-neutral design
Use cases, decision rules, prompt structures, quality criteria and review processes are documented independently of one vendor wherever practicable.
Reusable school assets
Schools receive agreed frameworks, templates, procedures and configuration guidance that staff can maintain and adapt within the applicable licence and contract.
Distributed capability
Leadership and staff development establishes more than one internal owner, reducing reliance on a single enthusiast, platform or external adviser.
How vendor lock-in is reduced
Core Education separates the educational logic from the delivery platform wherever technically possible. If a tool changes, the school retains its documented use cases, governance decisions, workflow logic and quality standards. Migration may still require technical work, and some platform-specific features cannot be transferred exactly, but the school does not have to rediscover the strategy from the beginning.
From one strategic briefing to sustained implementation.
Support may be limited to a presentation for a principal, executive team or whole staff, or extend into a contextualised roadmap, procedures, staff capability, custom configuration and ongoing advisory support. Detailed partnership models are presented once on the Architecture Services Overview rather than repeated here.