Research / Theme
Decision-Making Under Uncertainty
Institutional investment decisions rarely fail because of poor analysis. They fail because uncertainty is misframed, underestimated, or recognized too late.
In practice, decision-makers operate across incomplete information, conflicting signals, shifting regimes, and layered responsibilities. The most difficult moments are not those where options are clear, but those where it is unclear how to reason about the situation at all.
Under these conditions, traditional tools assume too much. They assume the right questions are already known, that risks can be isolated cleanly, and that decisions can be made sequentially. In reality, uncertainty compounds across portfolios, time horizons, governance constraints, and human behavior.
This research theme focuses on how investment reasoning breaks down under uncertainty — and how it can be strengthened. Rather than treating decisions as isolated events, we examine them as processes that unfold over time, shaped by assumptions, attention, and institutional context.
MARY addresses this directly through its multi-agent architecture. When signals conflict — macro suggesting caution while fundamental analysis finds compelling value — MARY surfaces the tension explicitly rather than averaging it away. The synthesis step identifies where agents agree, where they diverge, and which view carries more weight given current confidence levels. Decisions are informed by structured disagreement, not false consensus.