The governance layer for AI-assisted capital decisions.
Your team can research and draft capital decisions with any model, copilot, or platform. What no AI can do for you is prove each decision was permitted, keep it inside the mandate, and leave a record a regulator will accept. SigmaX does exactly that — every proposed capital action arrives mandate-checked, human-authorized, and audit-ready before it leaves the firm. SigmaX never moves the money itself.
Wealth firms, investment teams, fintech/advisor platforms, and institutional operations leaders.
Signals become emails, committee notes, and manual approvals while the firm still has to prove what is allowed before action.
A governed ActionIntent with mandate result, approval status, route status, impact record, and audit trail.
Faster review, fewer unsupported decisions, cleaner audit trail, and lower avoidable decision cost.
SigmaX does not trade, custody assets, move money, route customer capital, or replace human authorization. It governs the capital decision; the authorized human still makes it.
Proof before capital action.
Bring your own AI. SigmaX keeps every capital decision defensible.
Your team can research and draft with any model or platform it already trusts. What no AI can do for you is prove the decision was permitted, keep it inside the mandate, and leave a record that stands up to review. That governance layer is the product. SigmaX sits between whatever AI you use and any capital action — so speed never turns into a compliance liability.
Every capital-linked action — whoever or whatever proposed it — is checked against the account mandate and IPS before it can move. Anything outside what the account is allowed to do is held or blocked, not quietly passed through.
Deterministic Preflight and compliance checks run on every proposal, and a named human must authorize. No unsupported or undocumented decision reaches a route, and AI never holds capital authority.
Each decision preserves what changed, what was checked, who authorized it, and where it routed — an Impact Ledger and AuditWriter record you can hand to an auditor, a committee, or a regulator.
Any model, copilot, research platform, or agent can generate the proposed capital action and the context behind it.
Mandate, Preflight, evidence, cost, and route-eligibility checks run deterministically before anything can move.
SigmaX identifies who must approve, hold, or block. It never authorizes on the model’s behalf.
Impact Ledger and AuditWriter preserve what was proposed, what was checked, who authorized it, and where it routed.
Shipped today as the ActionIntent API. Fintech, advisor, and AI-agent platforms embed the governed action primitive — read-only and non-executing — so every proposed capital action returns a mandate result, required approver, route status, Impact Ledger reference, and audit trail.
AI analyzes. Deterministic systems validate. Humans authorize. Capital remains governed.
One governed API. Atlas onboards you to it.
Every business connects to the same governed surface: read the intelligence, submit a governed proposal — never move capital. And Atlas, as a read-only implementation engineer, walks your team through setup. What is live, what is built, and the boundary that makes it trustworthy are labeled on every capability.
Drift Radar is an entitlement-scoped, GET-only read API today. Fiduciary, decisions, and liquidity reads join it under one unified governed key as that plane rolls out. No authority, no writes.
Explore the API →Agents create at most a draft ActionIntent through the human-gated Agent Gateway — every response asserts no capital moved and execution off.
Explore the API →Atlas reads your setup read-only and names the exact next step. It never enters a credential, provisions, or authorizes — a human does.
Explore the API →AI proposes. Humans authorize. Capital remains governed.
Find your team. Walk a real decision, end to end.
SigmaX is built for firms and platforms with capital workflows where every proposed action needs evidence, mandate authority, human authorization, route status, and an audit trail before anything can leave the system — not for a retail trader looking for a stock pick. Pick your category and take a sample decision from inputs to a finalized ActionIntent.
Wealth firms and advisory teams
Portfolio drift, client constraints, tax context, and model changes create proposed actions that get resolved through scattered emails, memos, and manual approvals.
SigmaX turns the signal into an ActionIntent tied to the client Capital Twin, mandate, evidence, required approver, route status, and audit record.
Faster review, fewer unsupported recommendations, cleaner client-facing rationale, and a record the firm can defend later.
Investment teams and committees
Research and model signals produce possible allocation changes, but the committee must prove what is allowed, what evidence supports it, and who signed off.
SigmaX runs mandate checks, Preflight, counter-evidence, authorization requirements, paper or draft route status, and Impact Ledger records.
Less manual glue between research and committee action, stronger governance, and lower operational ambiguity.
Fintech, advisor, and AI-agent platforms
Platforms can generate signals or tasks, but they need a neutral permission object before anything can route to an execution, workflow, or human review system.
The ActionIntent API returns allowed, blocked, approval-required, paper-ready, or draft-handoff state with proof and audit references.
An embeddable governance primitive without becoming the broker, custodian, payment processor, or capital router.
Operations, risk, and compliance leaders
Approvals, evidence, exceptions, and route decisions are often reconstructed after the fact across tools and inboxes.
SigmaX keeps mandate version, approver, Preflight result, route status, impact, and AuditWriter record together from the start.
Cleaner audits, visible authority boundaries, fewer missing approvals, and a repeatable workflow for capital-linked decisions.
Each walkthrough is an illustrative sample. SigmaX analyzes and governs; a human authorizes; capital remains governed. It never trades, custodies assets, moves money, or routes customer capital.
The layer that turns signals into governed decision packets.
Firms have research, models, portfolios, notes, policies, and execution rails. The fragile step is the space between "something changed" and "what are we allowed to do next?" SigmaX turns that step into a governed decision packet: the proposed action, mandate check, Preflight, human authorization, route status, impact record, and audit trail. SigmaX calls that packet an ActionIntent.
AI analyzes. Deterministic systems validate. Humans authorize. Capital remains governed.
Any model, team, or platform drafts the capital action and the context behind it.
Mandate, Preflight, cost, and evidence checks run deterministically before anything can move.
SigmaX names who must approve, hold, or block. It never authorizes on your behalf.
The outcome is recorded as paper, an advisor task, a broker-handoff draft, or blocked.
What was decided, why, and by whom is preserved as a reviewable decision record.
The route is always one of four states -- paper, an advisor task, a broker-handoff draft, or blocked. SigmaX is not a broker, custodian, trading platform, or payment processor. No live capital authority is granted by AI or by the client portal.
How the chain works →Run the sample: company input to SigmaX evaluation to output packet to cost avoided.
Watch a company enter a capital problem, then change the customer, signal, and review cost. SigmaX turns the input into an ActionIntent, checks mandate and Preflight requirements, identifies the human authorization path, and produces the packet a company can use before anything outside SigmaX can act.
Company enters context
SigmaX finds the authority state
Advisors lose time turning drift, client constraints, and tax context into an approval-ready recommendation.
The proposed action is not treated as a recommendation or an order. SigmaX finds what changed, what the mandate allows, which proof is missing, who must review, and whether the route is blocked, held, paper-ready, or approval-required.
- 1Company enters contextA client Capital Twin moved outside the drift band after market movement and a recent deposit.
- 2Signal becomes ActionIntentCreate an ActionIntent to rebalance toward the mandate model without violating liquidity or tax constraints.
- 3Mandate and Preflight runClient IPS permits rebalance only when drift, tax exposure, liquidity, and suitability notes are checked.
- 4Human authorization is identifiedLead advisor plus compliance review for client-facing rationale
- 5Decision record and audit output are producedApproval required: advisor task with broker-handoff draft held until human authorization
Current tax-lot exposure, near-term cash needs, client notes, and model change rationale.
Recent client liquidity request could make immediate rebalance inappropriate.
Output packet the company receives
Approval required: advisor task with broker-handoff draft held until human authorization
Advisor, client file, and compliance archive
Lead advisor plus compliance review for client-facing rationale
Current tax-lot exposure, near-term cash needs, client notes, and model change rationale.
Mandate version, evidence class, route status, and reviewer are preserved.
Estimated labor cost avoided, measured against the full manual path (assemble + rework) minus the human review still done with SigmaX. Labor only; the rework share is an illustrative assumption, not a measured rate.
- - Less manual packet assembly before review.
- - Less rework from missing evidence or unclear authority.
- - Less audit reconstruction after the decision.
Simulation uses sample data only. No external action, no broker connection, no custody, no transfer, and no capital movement occurs from this website.
The sample shows one decision. Here is a year of them.
The estimate above is for a single ActionIntent. Change the inputs to your own book of business and the same method scales across a year of governed capital decisions. The labor it saves is the smaller half of the story; the larger half is the risk it takes off the table.
Sample inputs shown; edit any field to your firm. Labor only, and the rework share is an illustrative assumption, not a measured rate — so treat this as a sizing estimate, not a quote.
The decision that never lands is worth more than the hours.
Labor is the small, repeatable saving. The bigger one is the decision that should not happen — stopped before it does, because the route cannot open until authority is proven. These are the expensive failures: the compliance finding, the restitution, the remediation, the reputational hit.
never becomes a route. The mandate is checked before anything can move, so a rebalance or transfer the policy does not permit is stopped, not flagged after the fact.
is held, not assumed. Preflight names the evidence the decision requires up front; the route waits for it instead of proceeding on a gap.
cannot leave SigmaX. Nothing routes to paper, handoff, or a workflow until the named human authorization is on record.
does not exist here. Mandate version, evidence class, route status, and reviewer are written at decision time — not reassembled months later under examination.
SigmaX removes these structurally, not with reminders. AI analyzes. Deterministic systems validate. Humans authorize. Capital remains governed. No live capital authority, and SigmaX does not place trades or route customer capital — it produces the record a human authorizes against.
Sample data only. No external action, no broker connection, no custody, no transfer, and no capital movement occurs from this website.
Compliance isn’t a policy binder here. The software enforces it.
Every proposed capital action is checked — deterministically, before it can move — against the standards your obligations come from: your client mandates, suitability documentation, tax rules, firm limits, and trading conduct. What fails is blocked or held. What passes is documented. A named human authorizes. And all of it lands in a record built to withstand examination.
Every proposed action is recomputed against the signed governing document — allocation bands, allowed assets, position caps, prohibitions.
Sample: 71% − 9 pp = 62%, checked against the 45–65% band; a 15% position blocked by a 10% cap.
Recommendation documentation is assembled with the decision: rationale, alternatives considered, cost, and risk-profile alignment.
Sample: the rebalance file carries all four before the advisor can authorize.
Harvested lots are screened against the 30-day substantially-identical test before a replacement can be booked.
Sample: one replacement flagged and routed to the advisor — held, not passed through.
Concentration and exposure limits are enforced continuously; exceptions require sign-off, evidence, and an explicit expiry.
Sample: 6.2% vs a 5.0% issuer limit → a 48-hour, evidence-backed exception with a cure task.
Restricted names are screened out, and any desk handoff carries the firm’s best-execution requirements (FINRA 5310-style duties).
Sample: the committee’s broker-handoff draft ships with best-ex requirements attached.
Records are complete, append-only, attributable, time-stamped, and retrievable — the qualities examiners test record-keeping against.
Shown below: pull the record, read it, verify its integrity.
Configured per firm — these are the sample rule sets used in the walkthroughs above. Your mandates and policies become the machine-checked standard.
And when the examiner asks — the proof.
How each decision was made, the sources behind every fact, and who authorized it — pulled as a query, not reconstructed from inboxes. Sample data, real product surfaces.
Your compliance team — or your examiner, with your permission — pulls the record with a scoped, read-only query.
GET /v1/trigger-engine/audit?from=2026-01-01&intent=ai_demo_wealth_advisory Authorization: workspace-scoped key · read-only · least privilege → 200 OK · 8 entries · chain: intact (sample)
GET …/audit/export.csv
The complete record as a file your examiner opens in a spreadsheet — injection-safe, deterministic columns.
GET …/impact-ledger/export.csv
What each decision changed, exportable alongside the audit trail it reconciles against.
Behind the record: append-only storage enforced at the database level · named actors on every entry · provenance references and evidence classes (measured / derived / inferred) on every fact · workspace-scoped, least-privilege access · CSV and API export · recompute-on-read integrity verification.
Sample data shown; the query surfaces are the product’s real read-only audit APIs. Standards are referenced as what the data is checked against, configured to your firm — SigmaX is software, not legal advice or a certification. Your compliance team maps these controls to your regulatory obligations.
The buyer gets permission status, route status, and a record they can defend.
SigmaX is the capital-intelligence layer that makes decisions defensible. It tells a team whether a proposed capital-linked action is allowed, who must authorize it, where it may route, and what record proves the decision later. AI analyzes. Deterministic systems validate. Humans authorize. SigmaX does not move capital.
Positions, goals, mandates, constraints, source state, and evidence held as one decision context. SigmaX determines what is allowed to happen next on that Twin.
Research, counter-evidence, scenarios, drift, and provenance stay attached to the decision they support or challenge — every claim carries an evidence class and its limits.
Every outcome stays labeled as paper, advisor task, broker-handoff draft, or blocked. The status is a record for humans and systems; it is not execution.
SigmaX records what changed, who authorized, what was held or blocked, what route status existed, and the audit trail that proves the lifecycle.
Current public posture: no shadow-qualified opportunity is treated as customer-ready until evidence gates and governance review clear. No external action, live allocation, or customer-capital path is exposed here.
Eight offers. Three ways to buy the same proof-before-motion system.
Buy SigmaX as team software, an API for your own platform, or a private institutional implementation. Every path produces the same customer result: a proposed capital action is checked, authorized by a human, routed as paper/draft/blocked, and preserved as an audit-ready record.
Builder Sandbox - Decision Friction Audit
IC Memory Sprint - Private Pilot
Authority Team - Institutional OS - ActionIntent API
Run real paper or draft approval workflows for one team, on an annual platform.
Govern many Twins, mandates, approvers, impact records, and audit records.
Embed the governed action primitive in another platform or workflow.
Deploy a custom institutional authority layer privately.
Controlled availability for institutions wiring read-only decision intelligence into their stack. Arranged with the team under a supervised engagement.
Selective Access. Institutional Discipline.
Inquiries are reviewed individually for alignment with evidence-governed allocation discipline and long-term institutional fit. Past performance does not guarantee future results. Market risk includes possible loss of capital. Suitability, legal, compliance, and capital-governance review govern any capital-linked participation.
SigmaX provides capital decision intelligence, scenario analysis, and authority workflow software. SigmaX does not provide investment advice. SigmaX does not place trades, move funds, route customer capital, or manage client capital. Shadow/simulated/paper/backtested results are not live trading results. No result is a guarantee of future performance. Human approval is required before any capital-linked action.
Where SigmaX uses AI, it is governed as assistive software: truthful capability claims, human oversight, traceable records, disclosure of AI assistance, and no capital authority. SigmaX designs AI controls with reference to applicable securities obligations, FINRA existing-rule guidance, SEC AI-claims discipline, NIST AI RMF-style risk management, and EU AI Act transparency readiness. This is compliance-readiness design, not a claim that every deployment or jurisdiction-specific use has completed legal review.