PRD: M3 Portfolio Intelligence and Advisor
Problem
Raw holdings data does not tell users what actions are reasonable under current market conditions. Users need context-aware analysis tied to their own exposures and goals.
Objective
Deliver portfolio diagnostics, scenario analysis, and disciplined action plans grounded in user-specific data.
Success criteria
- Users receive exposure and concentration analysis they can understand quickly.
- The advisor can explain how macro and geopolitical events map to portfolio risk.
- Action plans are reviewable, assumption-aware, and non-executing by default.
Scope
- Portfolio exposure analytics
- Scenario and stress testing
- Personalized market and geopolitical briefings
- Research memos and thesis tracking
Out of scope
- Autonomous order execution
- Opaque strategy optimization
Requirements
- Separate facts, inferences, and recommendations in every output
- Track evidence and assumptions for each memo
- Provide user-readable risk and confidence summaries
- Support multiple horizons such as short-term risk and long-term planning
Risks
- Generic AI summaries with low signal
- False precision in scenarios
- Advice quality degradation from stale data
Delivery notes
The advisor should feel like a rigorous investment committee, not a prediction toy.