AI Trading Command and Control Interface Design
Nexus Trading: Agent Oversight and Execution Control
Product design study for a trading desk: fleet supervision, agent reasoning, pre-execution intervention and post-trade investigation.
Independent product design study. Sample data.
Nexus Trading explores how a trading desk supervises autonomous AI agents. I designed how a trader monitors the fleet, reviews pending actions and investigates executed trades.
My role: Product design, information architecture, interaction design and UI design.
Command Center
Monitor agent status, open positions and recent events, with individual controls and Kill All available in the same workspace.
Fleet status, position matrix, activity feed and global emergency control. Full interface.
Agent Control Panel
Review the agent's thesis, weighted signals and data freshness. Approve, reject or modify pending actions before auto-execution.
Agent reasoning, weighted signals, performance, pending decisions and open positions. Full interface.
Trade Blotter
Find trades by date, agent, instrument or P&L. Compare quantities, execution prices and fees, then export the journal to CSV.
Filtered trade journal with execution details, fees, P&L and CSV export. Full interface.
Agent Reasoning Replay
Reconstruct a decision through evaluation, risk check and execution, with recorded inputs and the market snapshot for the selected moment.
Decision timeline, recorded agent reasoning and the market snapshot at the selected moment. Full interface.
Design outcome
I connected fleet monitoring, decision review and post-trade investigation into a single supervision workflow. The central design principle is to keep the evidence behind an agent's action close to the controls for human intervention.
0 likes · 0 views · Oct 25, 2026
Product design study for a trading desk: fleet supervision, agent reasoning, pre-execution intervention and post-trade investigation.
