// For teams that run Foundry as their main data platform.
Platform control for Palantir Foundry,
inside your tenant.
Adrian turns your Foundry tenant into one ontology. When something goes wrong, you get an alert with the fix attached.
100% in your tenant
No new infrastructure
// 01 · The challenge
The more you build on Foundry, the harder it gets to know what it costs and how secure it is.
As projects multiply, waste and stale access build up, and whether they get caught depends on a few senior engineers noticing in time. What slips through is paid in cost, audit findings and senior time.
01 · Cost
Why is it so expensive, and what can we change?
Compute and storage grow every month. Some of it is work worth paying for and some of it is pipelines rebuilding what they already built. The invoice does not say which, let alone what to change.
02 · Security
Who has a role granted directly to them?
Access should come through groups. In practice roles get granted to individual people, in a hurry, on one resource. Across the tenant nobody has the list, until it becomes a problem.
// 02 · What Adrian does
Your tenant, modelled as an ontology.
Adrian reads Foundry's own metadata and keeps the tenant as an ontology in your Ontology: every resource, connection, schedule, principal, role, marking, cost and use.
What it helps you get right
More than twenty signals in four categories, watched continuously in the ontology. When one fires, the alert comes with its fix.
Architecture control
Security control
Pipeline health control
// 03 · How it works
From deployment to resolution in three steps.
01
Deploy
Adrian deploys into your existing Foundry tenant.
02
Model
It reads the platform's own metadata and builds an ontology that mirrors Foundry.
03
Resolved
Adrian watches the ontology continuously. When a signal fires, AI resolves it and opens a pull request.
// 04 · FAQ
Questions we get from platform teams.
Does any data leave our tenant?
No. Adrian runs entirely inside Foundry. Transforms, ontology, functions and front-end are deployed as native resources in your tenant.
What does resolution actually look like?
Each alert is tagged with how it gets resolved. Automatic: the fix is deterministic, one click applies it. AI resolves: an AI drafts the change, a person approves, it runs. Needs manual review: the fix is a judgement call, so the owner gets the alert with why it fired and a suggested fix, and decides.
Can I extend it?
Yes. Adrian's ontology is a normal Foundry ontology in your tenant: query it, link it to your own object types, build Workshop apps and automations on it.
// Get started
Ready to take control
of your tenant?
Pick a 30-minute slot.