The operating thesis

The firm remains the authority.

AI can widen inquiry, connect evidence, and prepare proposed work. The firm still determines what may be accessed, which policies apply, who may approve an outcome, and what may enter its established records workflow.

Five operating principles

Regulatory responsibility expressed through product design.

  1. 01

    Evidence before assertion.

    Material work should remain connected to its source, version, assumptions, and relevant context.

  2. 02

    Review before action.

    AI may assist analysis and preparation. Accountable professionals retain approval authority for consequential work.

  3. 03

    Records with continuity.

    Decisions, revisions, approvals, and supporting material should remain understandable and reviewable.

  4. 04

    Policy before autonomy.

    Firm-approved permissions, processing choices, and workflows define what an automated system may do.

  5. 05

    Responsibility remains human.

    Technology does not replace an adviser’s fiduciary, supervisory, legal, or compliance responsibilities.

The AI operating boundary

Extend professional capacity without obscuring professional responsibility.

ValaisOS is being designed so model-assisted work remains identifiable, reviewable, attributable, and subject to authorized approval.

  • AI involvement remains visible.
  • Sources and assumptions remain inspectable.
  • Material outputs pass through review.
  • Permissions constrain available actions.
  • Corrections preserve decision history.
  • Final responsibility remains with the firm and its authorized professionals.

Regulatory alignment

Design relevance—not a compliance conclusion.

These relationships describe where the ValaisOS design direction may support a firm’s own controls. Applicability and implementation remain firm-specific.

Regulatory concernValaisOS design direction
Books and recordsPreserve source context, versions, review, approval, and a clear promotion event into the firm’s designated workflow.
Compliance programsRepresent firm policy as a visible operating input rather than produce a generic compliance determination.
Marketing oversightSupport substantiation, balanced review, attributable source material, and retained approval evidence.
Privacy and safeguardingFavor minimized access, firm-governed processing choices, controlled data movement, and incident readiness.
Fiduciary responsibilityKeep professional judgment, material review, and accountable approval explicit.

Examination readiness in mind

Designed to make consequential work easier to reconstruct.

The product direction favors retrievable evidence, visible decision history, documented approvals, clear responsibility boundaries, and explicit handoff into firm-designated systems. Examination readiness depends on the firm’s complete implementation and operating procedures.

Public regulatory design register

A versioned view of the design commitments.

This register records public product direction. It is not a control certification, legal interpretation, or representation that a feature is generally available.

TopicDesign commitmentAuthority consideredStatus
Records authorityThe firm’s designated repository and records workflow remain authoritative.Advisers Act books-and-records requirements and firm policyPrinciple
Supervisory reviewAI-assisted material remains proposed until the appropriate review or approval occurs.Firm supervision and compliance policiesProduct direction
Marketing substantiationMaterial assertions should remain connected to evidence and review.Advisers Act marketing and recordkeeping requirementsProduct direction
Privacy and safeguardingAccess and processing should be minimized, controlled, and visible to the firm.Regulation S-P and firm information-security policyPrinciple
AI governanceThe firm defines permitted providers, processing boundaries, permissions, and approval paths.Firm policy and deployment-specific reviewProduct direction

Register version 1.0 · Reviewed July 24, 2026 · Revalidation required before product claims or deployment-specific control mappings.