Trust Operations Vocabulary

The Platform: r.Team is a Fiduciary Reasoning Platform with agents at the core. The map is one list: Reasoning Engine, the Bridge, Trust Intelligence, the Record, and a conversational layer. See the Platform.
Fiduciary Reasoning: The judgment work of trust administration. Distinct from trust accounting. See What is fiduciary reasoning?.
Fiduciary copilot: Software that works beside a trust officer: it reads, prepares, and cites. The named officer decides. Nothing happens until a person approves. Every decision goes to a tamper-evident record. See What is a fiduciary copilot?.
Family Trust Pod: Keep is offered by the trust company, under the trust company's name, as part of the service a Family Trust Pod provides. One Keep per Family Trust Account. See Keep.
Trust Administrator Pod: A Trust Administrator Pod attaches alongside Keep for the officer's work; the family never sees the officer's tools, and the officer never sees what the family keeps private. See What is a Trust Administrator Pod?.
Zero-data-retention AI for fiduciaries: Model providers do not keep client data for training; the firm controls retention and deletion; nothing is shared across trusts; only a fingerprint leaves the firm boundary. It does not mean the firm keeps no records: Keep writes every exchange to the trust's own record, and the Record is a tamper-evident log the firm is obliged to keep. Distinct from the website Privacy Policy. See Zero-data-retention AI for fiduciaries.
Trust Ontology with Securities Law: A structured map of the legal concepts that govern a trust and the securities-law and fiduciary standards a trust company operates under. The Reasoning Engine follows the map instead of guessing. See Trust Ontology with Securities Law.
Reasoning Engine: The shared context every agent reasons from. Known in engineering contexts as inference; on this site we name the capability by what operators experience.
Action Catalog: The firm's inventory of permitted agent actions. Agents do not invent unconstrained steps.
Review tiers: Four gates: clerical, structural, interpretive, and judgment. The agent cannot skip, batch past, or time out a gate.
The Bridge: One auditable doorway to systems of record. Read first. Write only after human approval.
Discovery agents: Agents that continuously review reports and operational data through the Bridge. Discovery proposes. The operator decides.
Trust Intelligence: Document intake and cited abstracts. Abstracts cite. They do not conclude. See Trust Document Intelligence.
Interpretation Governance: Firm-defined, versioned policies that decide what routes where. On the platform this lives in Trust Intelligence and the Reasoning Engine, not as a separate product.
The Record: A per-account hash chain of fiduciary events. Only a fingerprint is published.
Natural Language Interface: The conversational layer on the Platform. Questions stay inside the same gates as every other agent.
Peered Intelligence: Intelligence that works alongside your team at eye level: proposing and explaining, never overriding.
Trust Operations: The arithmetic and control side of the trust department. Distinct from — and complementary to — trust administration's document and judgment work.
Fiduciary Role: A governed responsibility in the trust relationship with authorities, duties, and approval thresholds.
Fee Schedule: The contracted rules for how fiduciary and administrative fees are calculated and applied.
Beneficiary: A person or entity entitled to present or future benefits under a trust or related structure.
Custodian: The institution that holds assets and produces position, transaction, and cash files. The Bridge connects to custodian feeds rather than replacing them.
NIL Intelligence Likeness: Name, Intelligence, Likeness — keep, download, and publish. You are not being replaced. You are being published. See the concept paper.
Neurosymbolic AI: Combining language understanding with explicit, verifiable rules. Language understanding alone guesses. Rules alone cannot read the instrument.
Inference: The engineering term for running a trained model. On this site we call this capability the Reasoning Engine.