Field guide · THE ESSENTIALS
What is an AI trading agent?
From a conversation to a plan that can take action. Understand the pieces, and your role in the loop.
Start with intent
An AI trading agent translates a request into a structured workflow. That workflow may gather market context, watch a condition or place an order through a supported venue. The useful distinction is what the agent is allowed to do after the conversation.
Research and execution are different steps
A research answer helps you think about a question. An execution plan needs explicit details: venue, instrument, size, entry condition and risk controls. In Mojo, you review a strategy card before the agent starts running it.
The conditions need a clock
Monitoring is not automatically continuous. Mojo’s documented scheduled checks can run as often as every 10 minutes; supported events can also wake an agent. A temporary move between checks may never become an observed trigger.
You remain responsible for the decision
An agent can misunderstand an instruction, receive incomplete data or fail to execute. Inspect the plan and its activity. Consider what happens if the service is unavailable, and understand which risk controls live on the exchange versus in the agent.
Product source: official Mojo documentation ↗ · Reviewed 7 October 2026.
← Back to the field guides
