Platform
Enterprise
Give a whole organisation one AI surface with the controls a security team will sign off on, and a bill you can actually understand.
Access follows your permissions
Retrieval is trimmed to what each user is allowed to see, so a document cannot leak out through an answer.
Spend you can cap
Per user and organisation wide limits, with cost visible per team rather than as one unexplained invoice.
Auditable by design
Every run records what was asked, which models answered, and which tools were touched.
How it works
- 01
Scope the rollout
Decide the teams, the data sources and the limits before anyone signs in.
- 02
Connect and clamp
Wire up your sources with permissions preserved and set spend ceilings.
- 03
Roll out and watch
Expand team by team with usage and cost visible throughout.
Try asking
What would it take to roll this out to two hundred people across four teams?
How do you stop one team's documents showing up in another team's answers?
Show me how spend is controlled and reported across an organisation.
Questions
- How is tenant isolation enforced?
- Every account is a separate tenant. Retrieval is filtered by permission at query time and filtered again before an answer is assembled, so a document outside your access cannot reach the prompt.
- Can we use our own model keys?
- Yes. Bring your own provider keys and usage runs on your accounts, with the keys sealed at rest and never returned to the client.
- How do we control cost?
- Spend gates operate per user and across the organisation, with hard ceilings that stop work before it exceeds a limit rather than reporting it afterwards.