From mandate requirements to a defensible wrapper decision — with the reasoning attached.

FinLattice is the internal desk tool for structuring private-credit, single-company equity and single-asset real-estate mandates. It generates wrapper candidates from declared requirements, explains every exclusion, and scores only what is actually feasible.

Pilot use cases: A1 · A2 · A3
Workflow: Define → Generate → Score
Every run: Versioned & reproducible

Operating principles

The workbench is opinionated by design: it refuses to produce a recommendation it cannot explain.

Outcome-first, wrapper-agnostic

Structuring starts from the mandate outcome. The wrapper is a conclusion, never an assumption.

Deterministic rules before inference

Hard rules run first and are reproducible. AI assistance is labelled and never decides a hard outcome.

UNKNOWN is never PASS

Missing facts stay unresolved with a recorded reason. Nothing is silently treated as zero or satisfied.

No silent disappearance

Every excluded candidate carries a reason code, a human explanation and its evidence status.

Score cannot rescue a hard failure

Scoring applies to feasible candidates only. Hard-failed structures are shown as non-scoreable.

Versioned and reproducible

Each run is pinned to a ruleset version with full lineage for committee review.

First pilot scope

Use case	Mandate shape	Primary structuring tension
A1	Private-credit loan / receivable	Origination capability, withholding leakage, custody rails
A2	Equity deal in one operating company	Concentration, investor eligibility, exit mechanics
A3	Single real-estate asset / SPV	Asset-level tax, local vehicle requirements, liquidity model

Synthetic demo data. Wrappers, rules, evidence and scores are fabricated for design review and are not legal, tax or regulatory advice.