On your network
Host the application and its stores in your environment. Local model endpoints are supported for inference.
Review every enabled service before treating a deployment as fully offline.
A platform for engineering decisions, with clear boundaries around your data, your people and your control systems.
Explore our approachPlus Two Sigma operates at Purdue Level 3 and above. It reads the data made available to it to help your engineers investigate issues, analyze performance and prepare work.
Your organization decides what it can reach. Approving a process recommendation records a decision; it does not change a setpoint or send a command to a DCS, PLC or safety instrumented system.
Application hosting and AI processing are separate choices. Agree both before connecting plant data.
Host the application and its stores in your environment. Local model endpoints are supported for inference.
Review every enabled service before treating a deployment as fully offline.
Plan a deployment in infrastructure your organization manages, with network access and storage policies set by your IT team.
Hosting, identity and operational requirements are scoped for each deployment.
Keep application data in your environment while using an external model service for selected workflows.
External processing can include prompts, retrieved context and tool results.
A local installation alone does not mean every feature runs locally.
Administrators manage accounts and permissions for agent use, connection management and signing decisions.
Proposed actions go through an approval or rejection step. The record includes the signed-in account and the decision.
Action decisions are written to an append-only, hash-chained audit log. Access to the log and its storage is managed in your deployment.
Infrastructure access, backup protection and log retention are part of the deployment review. Hash chaining supports integrity checks; it is not immutable storage.
These frameworks provide context for evaluating AI governance and security controls. The capabilities described here are not a statement of conformity or independent assurance.
An organizational management-system standard for responsible AI development and use, including governance, risk management and continual improvement.
An independent CPA examination of a service organization’s controls against applicable AICPA Trust Services Criteria. The result is an assurance report.
These references do not imply endorsement, certification, a completed audit or full alignment with either framework.
Use your security review to agree the data sources, hosting environment, external processing and account permissions for your deployment.