Security

Your institution’s knowledge.Inside boundaries you control.

Finpilot is designed to protect institutional investment information—from connected sources and user access through model requests, review and audit.

A quiet landscape framed by strong natural divisions
Security boundary
  • Institutional environmentControlled
  • Approved sourcesConnected
  • Permitted users and workflowsBy role
  • Review and auditRetained where supported

Every request crosses a defined boundary.Your institution stays in control.

Finpilot limits each request to the permitted task, context and provider policy—then returns the response to the institution for review.

  1. Connected sources

    Approved institutional information

  2. Institutional boundary

    Access, roles and permissions

  3. Permissioned request

    Only the context required for the task

  4. Approved provider

    Applicable policy and contract terms

  5. Response and review

    Sources, exceptions and human judgment

One question at a time.Answered plainly.

01

Is our data used to train models?

Customer data is used to deliver the contracted service and is not used for model training under Finpilot’s approved policy and applicable provider contracts.

Provider terms and platform policy govern model requests.

02

Is our information mixed with another institution’s?

Each institution’s records, documents and retrieval index are kept within its own boundary.

Institutional boundaries are part of the platform architecture.

03

Who can see what?

People see only what their roles and permissions allow. Workflow participation, review, approval and administration remain distinct.

Authentication, roles and permissions are checked on each request.

04

What happens when AI is uncertain or wrong?

Sources remain linked where available. Missing, conflicting or consequential information becomes a review item, and qualified people retain judgment.

Configured workflows determine what requires review and who approves it.

01Is our data used to train models?

Customer data is used to deliver the contracted service and is not used for model training under Finpilot’s approved policy and applicable provider contracts.

Provider terms and platform policy govern model requests.

02Is our information mixed with another institution’s?

Each institution’s records, documents and retrieval index are kept within its own boundary.

Institutional boundaries are part of the platform architecture.

03Who can see what?

People see only what their roles and permissions allow. Workflow participation, review, approval and administration remain distinct.

Authentication, roles and permissions are checked on each request.

04What happens when AI is uncertain or wrong?

Sources remain linked where available. Missing, conflicting or consequential information becomes a review item, and qualified people retain judgment.

Configured workflows determine what requires review and who approves it.

Four foundations.One controlled environment.

Independent assurance, institutional isolation, encryption and governed model use protect how information enters, moves through and leaves Finpilot.

  • Independent assurance

    SOC 2 Type II, with the report shared during diligence.

  • Institutional isolation

    Each institution’s information is kept within its own boundary.

  • Encrypted in transit and at rest

    Information is protected as it moves between systems and while it is stored.

  • Governed model use

    Customer data is not used for model training under Finpilot’s approved policy and applicable provider contracts.

Bring your security questions.We’ll review the evidence with your team.

When diligence begins, we’ll walk your team through the architecture, controls, data flow, provider terms and independent assurance—and share the appropriate documentation.