AI you can put in front of a security review, because you can state what it is allowed to do.
Governed AI execution with explicit operating modes, per-tenant policy and budget control.
The problem
What goes wrong without it
AI in an operational setting is only deployable if you can say what it may do, prove what it did, and cap what it costs. Most products answer none of those three.
AI Intel is the execution layer beneath every AI surface in the platform. Requests pass through policy, budget and grounding before reaching a provider, and the boundary is defined by risk class rather than by prompt.
Capabilities
What AI Intel does
Described from the current product build. Availability can depend on your deployment and plan.
- Per-tenant policy and per-user budget enforced before execution
- Tool authorisation boundary defined by risk class
- Provider routing by operating mode, with fallback
- Bring your own provider credentials where the plan allows
- Execution recorded for review
Operational workflow
How it is used
These are ordered stages of real use, not a feature list rearranged.
- 01Request
An operator intent arrives from any AI surface.
- 02Policy
Role and tenant entitlement are checked.
- 03Budget
The per-user limit is enforced before spend occurs.
- 04Ground
Live operational context is attached.
- 05Route
The request goes to the provider the mode selects.
Operating it
Integration, security and deployment
Integrations
- AI providers, per organisation
- Organisation-supplied credentials for BYO provider
Security & permissions
This is the module that makes the others reviewable. Every AI action is attributable to a user, bounded by an entitlement, capped by a budget, and constrained to tools its risk class permits. Refusal happens before execution, not as post-hoc filtering.
Security modelDeployment
- At least one provider must be configured before AI surfaces are usable
- Budgets and policy are set per organisation during onboarding
- BYO provider credentials are available on the plans that include it
Documentation
Where the detail lives
Documentation for this module is being prepared against the product. It is not published yet.
We publish documentation only once it has been checked against the product, rather than listing procedures nobody has run. The documentation home shows what is written, in review or planned.
Documentation homeRelated
Modules that share this model
These read from the same sites, personnel and incidents.
See AI Intel against your own sites
A working session using your operational context rather than a scripted tour. If your operation fits a plan available online, we will say so rather than route you through a sales process.