How to review why an AI trading agent traded
Review the instruction, the information available at the time, the proposed action and the execution result. DXAP records agent turns, including no-trade turns, and supports debriefs so you can ask why the agent acted or stayed out.
Published by DXAP, a product of DX Research Group. Updated September 13, 2026.
Start with the decision rather than the outcome
A profitable trade can come from weak reasoning, and a well-explained decision can lose money. Ask what evidence the agent used and how it related to the strategy. Review whether the explanation reflects what was known at the decision time rather than a story built around the later price.
Separate reasoning, permission and execution
The model’s explanation describes its proposed decision. Policy checks determine whether the action is allowed. Execution records show what actually happened. Compare these stages before concluding that a trade matched the instruction. An intended order and a filled position are not interchangeable.
Turn a finding into a clear change
If the agent misunderstood a preference, clarify the instruction. If an enforceable boundary is wrong, review the settings and any approval proposal. If the question is about an order or position, inspect the account record. This keeps a conversational explanation from silently becoming a settings change.
Put it into practice
Debrief prompt: “What information mattered for this decision? What evidence argued against it? Which instruction did you follow, and did a configured limit change the action? Distinguish what you proposed from what executed.”
Questions about this workflow
Can I inspect why the agent did nothing?
DXAP records no-trade turns as well as trading turns. Ask the agent what evidence or constraint led it to wait.
Does an explanation prove a strategy works?
An explanation makes a decision inspectable. It does not establish profitability; evaluate actual execution and costs separately.
Read the product details
Run your strategy with DXAP
DXAP is a hosted AI trading agent platform for Hyperliquid traders who want to describe a strategy in plain language, set its limits and inspect its decisions.
No subscription or separate model bill. DXAP charges 2.5 bps (0.025%) on executed agent volume; Hyperliquid fees and funding are separate. Alpha access requires a referral code. The minimum is $10 USDC per agent, with no user paper mode.
Trading can lose money. Examples describe workflows, not expected returns.