Clear answers. Deliberate boundaries.

A governed intelligence layer is software that sits between an AI model and any client-facing outcome in an advisory firm and applies the firm's own policy, evidence requirements, and approval path to the work while it is being done. ValaisOS is being designed as that layer for advisory firms. Full definition.

Fiduciary data custody is the principle that an advisory firm holds client information, decision reasoning, and supporting documents as entrusted property, received with a record of what arrived, held under the firm's own control, versioned so changes are attributable, and returned or produced intact on request. ValaisOS treats it as a design property. Full definition.

ValaisOS is being designed to support the records an advisory firm's own compliance program requires under the SEC regulatory framework for investment advisers, including books-and-records obligations under Rule 204-2 of the Investment Advisers Act of 1940, by keeping inputs, sources, generated output, human edits, and approval connected in a reviewable record. The firm determines which material is a required record; ValaisOS is being designed so the reasoning behind consequential work can be reconstructed from evidence rather than from memory.

Advisor AI governance is the set of firm-defined controls that determine how AI systems may be used in an investment advisory practice: which systems are approved, what client information each may access, which model providers are permitted, how sources are reviewed, how outputs are approved, how records are retained, and who holds supervisory sign-off. ValaisOS is being designed to express those controls as operating behavior. Full definition.

ValaisOS is pronounced va-LAY O-S: Valais sounds like valet without the final t, and rhymes with cafe; the final s is silent. It is an intelligence and operating environment being built for private-wealth professionals in an AI-native world. As artificial intelligence reshapes advice, client service, and firm operations, advisory practices need an enduring foundation that preserves discretion, evidence, professional review, and accountable authority. The name comes from Valais, a protected Alpine valley where paths, resources, cultures, and generations converge, an expression of the continuity and stewardship ValaisOS is designed to support.

ValaisOS is being designed as an intelligence and review layer over firm-approved information environments. The firm continues to determine its authoritative repository and records workflow.

No. The product direction is to work with firm-approved systems rather than displace them. Each firm determines which platforms remain authoritative for client, document, and records activity.

No. ValaisOS is being designed for consequential professional work in which evidence, firm policy, review, and accountable authority remain connected to the intelligence being used.

The firm does. The firm determines permitted information, processing choices, approval paths, authorized actions, and the systems into which approved work may be promoted.

A governed intelligence layer is software that sits between an AI model and any client-facing outcome in an advisory firm and applies the firm's own policy, evidence requirements, and approval path to the work while it is being done. In a governed intelligence layer, AI may inspect, interpret, and propose; a licensed professional reviews and approves; and only approved work enters the firm's systems of record. ValaisOS is being designed as a governed intelligence layer for advisory firms, so accountability stays with the licensed professional. Read the glossary entry.

Governed AI is not size-dependent. Supervision, recordkeeping, and fiduciary obligations apply to a two-person RIA as fully as to a national firm. ValaisOS is being designed for advisory firms of any size, on the premise that governance belongs to the structure of the work rather than to firm headcount.

The product direction supports firm-governed choices about permitted providers, credentials, execution modes, data classes, and outbound boundaries. Availability will depend on the approved deployment.

Local processing is part of the design direction for appropriate workflows, alongside firm-approved remote options. Local execution is a control choice, not a compliance conclusion, and it still depends on endpoint, retention, security, and operating controls.

No universal claim is made. Data movement depends on the firm’s approved deployment and processing mode. ValaisOS is being designed to make relevant processing and outbound boundaries visible rather than imply that every workflow is local.

ValaisOS is being designed so model-assisted work remains distinguishable from approved firm work, with relevant sources, assumptions, processing context, and review state available to the authorized professional.

AI compliance monitoring reviews communications and activity after they occur to surface exceptions. Governed AI execution applies firm policy before and during the work, so output is produced inside the firm's controls rather than inspected only after the fact. ValaisOS is being designed as a governed execution environment; monitoring tools address a different layer of an RIA compliance program, and the two can operate together.

Product workflows are being designed to keep sources, assumptions, versions, review, and approval connected so a material outcome can be understood and reconstructed.

The regulation-first direction favors visible stale-source, policy-mismatch, conflict, and failure states. A failed or outdated workflow should not silently turn unreviewed output into authoritative work.

No. ValaisOS is being designed to treat AI output as proposed work. The firm determines classification, approval, retention, and promotion into its established records workflow.

The design direction treats promotion as an explicit event: authorized work moves to a firm-designated destination with its approval and resulting destination context preserved. Availability depends on the approved deployment.

The firm determines which material is a required record and which system is authoritative. ValaisOS is being designed to support explicit review and promotion into the firm’s established records workflow rather than make that legal determination itself.

That is the intended direction. Access, processing, review, approval, and promotion should follow a firm’s approved policy profile rather than a generic compliance conclusion.

No technology makes a firm compliant by itself. ValaisOS is being designed to operate within firm-approved policies, systems, configurations, and supervisory procedures.

Public design commitments and future control mappings will be versioned and revalidated before they support a product claim or deployment decision. Each firm remains responsible for interpreting changes with its own legal and compliance professionals.

A defensible AI policy defines controls before capability: which systems are approved, what client information each may access, how sources are reviewed, how records are retained, and who holds supervisory sign-off. ValaisOS is being designed to express those firm-defined controls as operating behavior, so an AI-assisted workflow follows the firm's approved-system, information-scope, and review requirements rather than a generic default.

Not yet. Product details and access remain selective while the product and review standard are being refined. Private-preview requests are considered individually.

A name and work email are enough to begin. Optional professional context can help make a future conversation relevant, but confidential client information should not be submitted through the public form.