evaluate_fca_handbook_applicability
Identify which FCA Handbook rules apply to a specific firm, product, or service, providing compliance reasoning for authorisation, permissions, and obligations.
Instructions
Evaluate which FCA Handbook entries apply to an entity, via the Metis FCA Handbook AI Harness.
Calls a live compliance reasoning run, billed to the configured account. Use for questions about FCA authorisation, permissions, or obligations for a specific firm/product/service.
CRITICAL: This call takes 90+ seconds to complete and streams progress messages. You MUST display each progress message to the user as it arrives. Do NOT wait silently for the result. Silently waiting makes the user think the tool is broken. The progress messages are not filler — they contain essential detail about what the Harness is doing (which reasoning node is running, what it found, etc.). Your response MUST actively incorporate and relay each message, not just the final result.
Before calling: check whether you already have (from this conversation, documents you were given, or other tools) grounded answers to these six things — the specific compliance question, the product/service, who's providing it (platform/adviser/bank/etc.), its key features, the target market (retail/institutional/professional), and what data it handles. If you are missing more than one or two, ask the user for them first rather than calling with thin input. Every call is billed to the account at a flat rate regardless of input quality, so a vague call followed by a refinement round costs twice what one good call would have.
The result includes refinement_suggestions — gaps the Harness couldn't resolve from user_input alone, typically subtler than the six basics above (e.g. a regulatory edge case, not a missing fact you could have just asked for). If you already have grounded information addressing one (from this conversation, documents you were given, or other tools you've called), fold it into a new user_input and call again yourself, rather than just relaying the suggestion to the user as a question. Do not speculate or infer plausible-sounding detail you do not actually have to fill a gap — that reintroduces the hallucination risk this Harness exists to avoid, one level up. Only ask the user for whatever's left that you genuinely do not know.
Args: user_input: Everything together as one piece of text (up to 5000 characters) — the specific compliance question, the product/service, who's providing it, its key features, the target market, and what data it handles. See "Before calling" above for why all six matter. analysis_mode: 'quick' (default, ~60-120 seconds) for a fast pass, or 'full' (longer) for detailed conditional reasoning — conditions, interactions between rules, and second-order implications. Ask the user which they want if it is not obvious; default to 'quick'.
Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| user_input | Yes | ||
| analysis_mode | No | quick |