Safe AI use, taught at every level.
Most AI incidents start with an employee pasting the wrong thing into the wrong tool. Each level teaches the rules that match what participants can now do.
| Programme | What participants can do | Security rules taught |
|---|---|---|
| AI Essentials | Use approved AI tools for daily work | What data stays out of prompts, approved tools versus consumer accounts, checking outputs before sharing |
| AI Practitioner | Share workspaces, connect email and drives, schedule tasks | What a shared workspace exposes, connector permissions, approval before anything leaves the company |
| AI Builder | Build tools that touch data and systems | API keys and secrets, least access, prompt injection, when a tool must move to IT |
| AI for Leaders | Set policy for the company | Guardrails, approvals, data residency, accountability for AI-assisted decisions |
How the Academy handles client data.
The client's tools
Sessions run on the client's own licences and accounts, under its data rules.
Nothing kept
Client documents used in exercises stay in the client's environment. Nothing is stored on Astra Via infrastructure.
IT and security involved
The client's IT and security teams see the programme before it runs, and can add their own rules.
When Astra Via builds.
- Data stays in region. The client's regional cloud, or on premises
- The client's tenant and keys. When Astra Via leaves, access ends
- Every action logged. Who or what did it, and which model was used
- Human approval by design. Customers, money and contracts gated behind a named person
- Reversible. The client owns the code, the prompts and the documentation