Idea Ripper // ArchiveCards: 1555 [Live] | Updated: 2026-10-14
Card #c21IR // #c21Synthesis

The architecture has four layers, illustrated in Figure 11.1 . Evidence and assurance flow upward from infrastructure and behavior to risk and accountability, while policy, limits, permissions, and escalation rules flow downward. The layers are interdependent rather than sequential maturity levels.

— In our words · Source: Large Language Models in Finance (Miquel Noguer i Alonso)

[Use when]

When designing or auditing a financial LLM system, verify all four layers agree — infrastructure, behavior, risk, accountability — instead of certifying the model alone.

[Why it matters]

Controls fail at layer boundaries, not inside layers: a calibrated model fed stale data, or fresh data flowing through unpermissioned tools, breaks the system even though each layer 'works.' The four layers are the checklist for where to look when something goes wrong.

Related principles

#c20Large Language Models in finance are best understood as part of a pipeline rather than as a point solution. Raw documents, time series, and…When a team frames an LLM project as 'pick the best model,' reframe the plan around the pipeline: curation,…

More from Large Language Models in Finance

#c13Prediction and agency should not be conflated. A credit model can estimate the probability that a borrower breaches a covenant. An agentic…Use when a team proposes a stronger model as the fix for an agent that lacks memory, permissioning, or an… #c14The technically and institutionally mature objective is bounded autonomy : the agent can plan, retrieve, calculate, and propose actions…Use when deciding what an AI agent is permitted to do in production: which actions stay read-only, which need… #c15This principle prevents a common category error. A model that says it will respect a mandate has not yet respected the mandate. A model…Use when a vendor demo, benchmark score, or red-team chat log is being offered as proof that an agent is safe…