One Control Layer for Data, Workflows and Agents
Registry, permissions, vaults, memory, approvals, evidence and model choice, defined once for data, workflows and agents alike. We do not govern to constrain. We govern so we can deploy more with confidence.
What the Estate Is Governed By
Registry
Every agent in the estate, who owns it, what it is for, when it last ran. No shadow agents.
Permissions
Which data an agent sees, which NGT Tools and NGT Skills it can reach, which systems it can write to.
Vaults
Credentials issued centrally, scoped per role, rotated centrally, never sitting in a prompt or a script. Revoke the agent and you revoke its access everywhere.
Memory
What agents retain across runs and share with each other. Scoped, inspectable, isolated where it has to be.
Approvals
Which actions wait for a human signature, set once rather than agent by agent.
Evidence
Every run logged, traceable, replayable. The answer to what an agent was allowed to do, who decided that, and what it did.
Model
You choose the model: NGT Auto or any AI lab's LLM. Swap it without rebuilding the estate.
The Harness Is What Makes an Agent Deployable
The Foundation the Control Layer Stands On
ISO 27001
Certified information security management, audited every year.
NIST 800-53
Security controls aligned to the NIST framework across the platform.
DORA
Operational resilience obligations for financial entities, built into how the platform is run.
GDPR
Personal data handled under the regulation, with data processing agreements for every client.
Single tenant
Every client runs in its own isolated environment. Methodology carries across deployments; data does not.
Sovereign
Data stays in the region you choose, under your keys, with full lineage of where it came from and where it went.
See the control layer on your estate