Institutional-grade analytics — governed, local, signed.

An analyst poses a question about holdings exposure. The platform computes the answer deterministically — tracing through fund structures, checking documents, surfacing relevant passages — then assembles a committee-ready memo. A named portfolio manager reviews and approves before the memo leaves the system.

Premium quantitative advisory — risk diagnostics, factor exposures, performance attribution, stress scenarios, regime and anomaly signals — computed for every client portfolio automatically. The AI translates the numbers into a personalised, house-voice narrative. A portfolio manager signs every piece before it goes out.

Every proposed order is checked before it reaches the market — against the fund's own mandate, against exclusion lists (including via the issuer's parent company), and against concentration limits. Anything flagged pauses for a named compliance officer to approve, override with a written reason, or reject. A rejected order is enforced fail-closed at the OMS: a system outage never silently passes an order through.

Takes a private client's documents — ID, financials, risk questionnaire, source-of-wealth letter — and walks through profile extraction, KYC/AML screening, MiFID II classification, risk-band assignment, and model portfolio matching. Produces a signed Geeignetheitserklärung. Everything runs on the firm's own hardware; no client data leaves the premises.

Points at the firm's existing client folders — however they were filed over the years — and produces a per-client compliance view: which required documents are present and valid, which are expiring, which are wrong, and which are missing. A deterministic rulebook makes every status decision; the AI only classifies documents. No file is ever moved or renamed.

An inbound fund document — prospectus, KID, or factsheet — enters the desk. The platform extracts all regulated PRIIPs RTS fields, validates them deterministically, lets a compliance officer verify any flagged field, drafts the KID in the firm's house voice, and requires a second compliance sign-off before rendering the final versioned PDF. Every field's lineage is audited.

Every model the firm runs is registered. A model that fails independent pre-deployment validation is blocked from going live. Once in production, each model is monitored for drift against a statistical bright-line computed by code — not by AI. Every status change is logged immutably. A one-click governance pack exports the full picture for regulators or auditors.

Systematic trading strategies are validated out-of-sample before deployment. Live strategies run under hard risk limits — max drawdown, leverage, position size. An armed kill switch, controlled by code, halts trading on limit breach. Every trade, parameter change, and halt is logged.

Every day, vendors deliver the same security at different prices. The platform compares feeds, decides by code whether the gap is material, identifies the root cause — most often a vendor not yet applying today's dividend — computes the NAV impact, and parks the case at a four-eyes gate for human approval. Every figure shown is computed by the engine; the AI only writes the explanatory note.

Generates personalised, compliant client letters triggered by portfolio events: a drawdown breach, a quarterly cycle, or a client question. Mandatory regulatory disclosures are assembled deterministically and cannot be omitted. The AI drafts in the advisor's house voice. The advisor approves before anything is sent.