Subprocessors
Subprocessor Notice
Last updated: 24 April 2026
This notice identifies the categories of third-party service providers Halo Education may use to deliver the service. Because Halo is designed to minimise central collection of student information — identifiable student records are not sent to AI — the data disclosed to subprocessors is generally limited to service metadata, account data, and de-identified, pseudonymised, or aggregate AI-processing context where relevant.
The current list of named providers is available on request at privacy@haloeducation.app. Enterprise customers receive the current list as part of their onboarding documentation.
1. Hosting and application delivery
Purpose: hosting frontend and backend services, application delivery, infrastructure operations, edge network, DNS, DDoS protection.
Data categories: service requests, limited operational metadata, deployment and infrastructure logs, account-related service traffic where relevant.
Halo is a school-hosted platform: each school deploys Halo inside their own approved cloud tenant and selects the region for that tenant. Australian schools host the whole application — including AI agent processing — within Australia.
2. AI model and inference providers
Purpose: generation of AI-assisted outputs, structured drafting, language workflows.
Data categories: de-identified evidence packets, pseudonymised context, aggregate metrics, prompt and response payloads needed to provide the feature.
Approved providers include major large language model platforms with enterprise privacy terms. The specific provider currently in use is disclosed on request and to enterprise customers. We do not permit approved providers to use customer data to train their own models.
3. Authentication providers
Purpose: optional sign-in and identity management.
Data categories: name, email address, account identifiers, authentication metadata.
Approved providers include Google, Microsoft Entra ID, and Apple Sign In. Enterprise customers may federate their own identity provider through the same standards (OIDC / SAML).
4. Error monitoring and service diagnostics
Purpose: detecting faults, debugging operational issues, maintaining service reliability.
Data categories: technical error information, service diagnostics, limited metadata needed to investigate faults.
Personally-identifying information is configured out of diagnostic capture where practicable. Sensitive student information should not appear in diagnostics under normal use.
5. Transactional email and support communications
Purpose: account emails, support correspondence, operational notices.
Data categories: contact details, correspondence content, support case metadata.
6. Enterprise deployment providers
Purpose: school-level or tenant-specific deployments.
Data categories: enterprise account and deployment information, governance and operational metadata, any other information processed under the enterprise deployment agreement.
Enterprise customers may require Halo to deploy inside their own approved cloud environment (for example, their existing Microsoft Azure tenancy). In that configuration the hosting subprocessor for enterprise workloads is the customer's own platform vendor, not a Halo-chosen provider.
7. Where data is processed
Because Halo is school-hosted, the school decides where data is processed — Halo deploys into the school's own cloud tenant in the region the school selects. For Australian schools, the entire application and its AI agents run within Australia. Specific countries and providers for any centrally-administered surfaces are disclosed in the named subprocessor list available on request and to enterprise customers.
8. Updates
We may update this notice from time to time as our infrastructure or providers change. Enterprise customers are notified of material subprocessor changes in line with their contract.
9. Questions
Questions about subprocessors — including requests for the current named list — can be sent to privacy@haloeducation.app.