MIA is a talking, listening wealth avatar backed by the Financial Twin— a living, consent-fed model of a customer's whole financial life across every bank and AA-linked account. It doesn't just recommend. It explains, tags suitability, and logs every decision for the regulator.
A voice-and-video avatar (Azure) fronts a FastAPI orchestrator that unifies accounts, holdings, goals and behaviour into one Twin, then reasons over it with an engine stack: surplus, scenario simulation, behaviour, life-events, suitability — one brain, three surfaces (customer app, this console, the avatar).
Every recommendation and every MIA answer carries a visible suitability tag (suitable / needs review / not suitable) and a structured rationale — inputs, assumptions, reasoning, risk disclosure — written to an immutable audit log. Not a settings toggle: the default.
Customer 360 · Recommendation feed · Audit trail · RM co-pilot — all synced in real time.