From natural language to enforceable mandate

Question answered: Do I need to learn a policy DSL? How does a user intent become a structured mandate?

You describe what you want in natural language. delta turns it into a structured, enforceable mandate.

A user might say: “I want a green sweater, size medium, 100% wool, max $100, from an approved merchant, delivered by Friday.” That sentence becomes a mandate with typed constraints — value bound on price, attribute matches on color and size and material, counterparty restriction on merchant, temporal condition on delivery — each grounded in real categories and evidence attributes from delta’s product taxonomy.

The taxonomy is what makes this work. It maps natural-language concepts to the specific, verifiable attributes the policy engine can check. “Green” resolves to a color attribute in the taxonomy. “100% wool” resolves to a material attribute. “Approved merchant” resolves to a counterparty restriction. Without the taxonomy, natural language is just text; with it, every word maps to a typed field the evaluator can test against extracted evidence.

The compilation happens via the MCP. An agent calls the language guide, searches the taxonomy for the relevant product categories and their evidence attributes, drafts the policy, and validates it with the policy engine compiler. If the compiler accepts it, the policy is stored and ready to be referenced by a signed intent. If it rejects it, the agent gets a compile error with a line and column — same feedback loop as any compiler.

The user never sees the DSL. They see the natural-language description and the structured mandate it produced. The gap between the two is the taxonomy and the compiler.