Research Agent
Synthesizes filings, transcripts, market data, and research notes into investable insight with citations.
AI Native
Spitfire embeds AI into the operating system itself through an Intent → Context → Reasoning → Simulation → Control → Action → Evidence → Learning loop.
Portfolio Agent
Translates investment views into sizing, rebalancing, hedge proposals, and scenario-aware alternatives.
Risk Agent
Explains risk movements, concentration, and stress outcomes with explicit drivers and controls.
Execution Agent
Supports liquidity assessment, slicing, venue selection, and execution quality interpretation.
Operations Agent
Investigates breaks, assembles evidence, and drafts operational resolution actions.
Compliance + Model Guardian
Enforces policy controls, approval flows, model inventory, drift monitoring, and release discipline.
AI Action Governance Framework
Spitfire classifies AI-supported actions by operational risk and maps each class to explicit approval and control requirements.
| Action Category | Control Standard |
|---|---|
| Informational and Explanatory Outputs | No pre-approval required |
| Recommendations and Draft Preparation | Portfolio Manager or Trader approval required |
| Execution Actions | Policy validation and explicit human authorization required |
| Model or Valuation Overrides | Dual authorization and immutable audit evidence required |
| Learning and Model Adaptation | Model governance and validation workflow required |