Business details
- Entity: AiBorz (sole trader — Alborz Fallah)
- Location: Brisbane, Australia
- Contact: [email protected]
- ABN: To be published upon registration
Data handling principles
- Client-owned systems. AiBorz prefers to work inside client-owned Microsoft 365, Google Workspace, OpenAI, cloud and SaaS accounts where practical. The client retains administrative control.
- Least privilege. AI workflows are granted only the minimum access needed for the approved task.
- Data minimisation. Only the data required to complete a workflow step is sent to AI models.
- No training on client data. AiBorz uses business-grade AI products where data is not used for model training by default. Where exceptions apply, they are disclosed and approved before use.
- Data storage. Client documents, prompts and workflow configurations are stored in client-owned systems wherever possible. AiBorz retains only the minimum information needed to operate and improve the service.
- Australian guidance. AiBorz designs workflows consistent with OAIC privacy guidance, ACSC small-business cyber recommendations and National AI Centre governance recommendations.
Credential handling
- Client credentials, API keys and access tokens are stored in client-owned password managers or secure credential stores where practical.
- AiBorz does not store client passwords, API keys or access credentials in plain text, shared documents or source code.
- Access to client systems is time-limited, scoped and documented.
Incident response
If an AI workflow produces incorrect, misleading or harmful output — or if protected data is accidentally exposed — the incident process is:
- Stop: pause the affected workflow immediately.
- Notify: inform the client's AI owner within 24 hours.
- Document: record what happened, what data was involved and what corrective action was taken.
- Review: determine whether the workflow, data rules, training or controls need updating.
- Report: provide a written summary to the client including root cause, impact and remediation.
Subprocessors
AiBorz may use the following in delivery of services:
- AI model providers: OpenAI, Anthropic, Google, DeepSeek or other models selected by task, risk and data requirements. Business-grade accounts are used where client data is involved.
- Automation platforms: Make, Zapier, n8n, Power Automate or equivalent — selected based on client system compatibility and security posture.
- Cloud infrastructure: Railway, Vercel or equivalent for hosting AiBorz tools and APIs.
- Communication: email via the client's existing provider or a configured SMTP service.
Specific subprocessors are confirmed during the AI Readiness & Risk Audit for each client engagement.
Offboarding
If a client discontinues AiBorz services:
- Client-owned systems and accounts remain with the client.
- Workflow documentation, prompts, configurations and operating procedures are handed over in a documented format.
- AiBorz removes its access to client systems within an agreed timeframe.
- Any AiBorz-retained client data is deleted within 30 days unless otherwise agreed or required by law.
For the full governance model, see the Security & Governance page. For the operational launch approach, see the Launch Protocol.