When AI touches the process.

Why would I document that AI was involved at all?

Because the record can work in your favor, not against you. Picture the alternative: a model helped screen a position, the committee signed off, and six months later no one can show what the model flagged or that a person checked it. That gap, where AI was used and nothing explains how it factored in, is the harder thing to account for later.

Isn't “the investment committee decided, like always” enough?

That describes the outcome, not the process, and it is incomplete if AI touched the analysis. A signed, timestamped record lets the firm point to a documented explanation made at the time, rather than reconstructing the process from memory months later.

Won't a written record just be used against me?

This is the real worry: that a written record of the risks and the dissent gives someone something to point to. But the weaker position is using AI and keeping nothing, which can look like the risk was never considered at all. A record that shows the firm saw the risk, weighed it, and had a licensed person decide anyway is a documented exercise of judgment. What to keep, and how, is a determination for the firm and its counsel.

What is actually in the record?

The decision, the risks that were surfaced, a confidence figure (a measure of how strongly the roles converged, computed from a published, deterministic formula, not a probability of being correct), and the preserved dissent attributed to the role that raised it. Each record carries a SHA-256 fingerprint of its contents, so two copies can be compared for an exact match. It is a signed, timestamped decision record, not a tamper-proof seal, and the adviser owns the decision. The formula and method are on /methodology.

Is this investment advice, or a compliance guarantee?

Neither. Teranode is decision-support infrastructure. It does not make recommendations to your clients and it does not certify compliance. It produces a structured record of a supervised reasoning process that your firm's licensed people review and sign. Whether any disclosure is required is a determination for your firm and its counsel.

Where the rules are heading.

Documenting how AI reaches a decision is moving from good practice toward regulatory expectation, and not only in the US. Across major financial jurisdictions, the same theme is emerging: if AI touched a regulated decision, a firm should be able to evidence and reconstruct it. Almost none of this is new AI-specific law; regulators are largely applying existing books-and-records, supervision, and model-risk rules to AI. In most jurisdictions this remains an emerging expectation, not a specific legal requirement.

Of the jurisdictions below, only the US rules apply to US registered investment advisers. The others are included to show the direction of travel, not because they bind a US firm.

• United States (binds US advisers).The SEC's 2026 examination priorities (released November 2025) name automated investment tools and AI technologies as a focus area. In that section the Division of Examinations said it will assess “whether firms have implemented adequate policies and procedures to monitor and/or supervise their use of AI technologies,” and will “review for accuracy registrant representations regarding their AI capabilities.” The priorities are a staff statement that, by their own terms, have “no legal force or effect.” Separately, the firm and its counsel apply advisers' existing obligations to AI-assisted activity: substantiation of marketing claims, books and records, and keeping the basis for advice under the fiduciary duty. There is no SEC rule that specifically requires advisers to document how AI reaches a decision; the expectation is a direction of travel, not a current mandate. A focused read of the 2026 exam priorities on AI is at /exam-readiness, and ongoing coverage is in the SEC AI Brief.
• Canada (does not bind US advisers). OSFI Guideline E-23 will bring AI/ML models under formal model-risk governance: enterprise model inventory, lifecycle documentation, independent validation. Finalized September 2025; effective May 1, 2027.
• European Union (does not bind US advisers). The EU AI Act framework is in force, but its high-risk obligations for finance (logging and technical documentation) apply only to credit scoring and insurance pricing, not to advisory or portfolio AI. The compliance dates for those obligations are still being finalized and are expected to defer toward late 2027 under pending proposals.
• United Kingdom (does not bind US advisers). The FCA governs AI through existing rules: the Consumer Duty, the Senior Managers regime, and systems-and-controls recordkeeping. In force.
• Australia (does not bind US advisers).ASIC applies its “efficiently, honestly and fairly” duty to AI, and APRA's CPS 230 operational-risk standard treats AI within operational and technology risk management. In force.
• Singapore (does not bind US advisers). MAS's proposed AI Risk Management Guidelines would require an AI inventory, risk scoring, and lifecycle documentation across a firm's AI systems. Proposed; in consultation, not in force.

The direction of travel is broadly similar across these jurisdictions, though the specifics (what is in force, what it covers, and when) vary, and most of these regimes do not bind US advisers. Teranode's decision record is designed to help firms capture a contemporaneous, documented basis for AI-assisted decisions. Teranode does not make any firm compliant or ready for an examination. What any specific rule requires, and whether any rule requires a firm to document how AI reaches a decision, is a determination for your firm and its counsel.

The founding pilot.

We are opening a small number of founding design-partner slots. A pilot runs 30 to 60 days and sits alongside your existing supervision; it does not replace any part of it. We scope it on a short call, run it on one of your real decision types, and your firm keeps every record it produces.

• What the firm receives. Access to run the Council on its own decision types, the decision records those runs produce, and a direct line to the founder.
• What is expected. Structured feedback on real workflows, so the routing and the record format improve against how your firm actually works.
• Price. Terms are set per firm and discussed on the call. Nothing here is an offer or a commitment of price.

Teranode produces records; your firm remains solely responsible for its supervisory system and every supervisory determination.