Discover
Inventory models, applications, agents, datasets, knowledge sources and tool endpoints; identify unregistered use.
Sovereign trust
Control AI risk across the full lifecycle.
Discover AI systems, inspect their supply chain, test behaviour, enforce runtime policy and retain evidence inside the operational boundary you define.
Status: Roadmap. Availability, release terms and supported deployment conditions will be published as dated capability records before the status changes.
Core capabilities
Designed around accountable organisational work.
Exact coverage, deployment requirements and limitations are documented for the selected version and use case.
Inspect
Examine model artefacts, repositories, dependencies and tool integrations for unsafe behaviour and supply-chain risk.
Validate
Test models and applications for prompt injection, jailbreaks, data extraction, privacy failures and unsafe tool use.
Protect
Apply policy to prompts, retrieved content, responses, code, URLs and tool actions; allow, block, redact or escalate.
Prove
Retain policies, evaluations, decisions, events and reports so governance and security teams can review the lifecycle.
Product boundary
A trust plane across the platform
AI Guard spans applications, models, assistants, tools and data interfaces. Framework mappings can organise evidence, but are not certifications or legal guarantees.
Adoption outcome
Understand exposure before expanding AI
Begin with an inventory and defined risk scenarios, then establish evaluation, release and runtime controls appropriate to the use case.
Sovereignty matrix
Evaluate the deployment across six dimensions.
Product-specific answers are established through the deployment architecture and dated evidence record.
DataOperationalTechnologyJurisdictionKnowledgeExit
Review the trust model
Connected capabilities
See how this responsibility fits the platform.
A practical first step
Understand exposure before expanding AI
Begin with an inventory and defined risk scenarios, then establish evaluation, release and runtime controls appropriate to the use case.