A clear ask
Build a systematic trading agent, not a source of return promises.
Case study · Trading agent
LiveAn agent built around its guardrails.
For a private investor in Montréal. The system is live, with a deterministic risk layer between decisions and execution.
Case study · Trading agent
Discipline before execution.
Private access
A private web interface supports conversations with the agent from a phone.
Controlled execution
The execution layer accepts only risk-approved position sizes.
Case study · Trading agent
How the agent works
The conversation interface receives the request; it cannot authorize risk.
Illustrative workflow. Select a step to explore it.
See system details
Case study · Trading agent
Four layers. Separate responsibilities.
The conversational interface receives questions. It does not decide whether a trade is allowed.
Case study · Trading agent
Prove it small before scaling.
01 / Backtest
Test with costs: positive expectancy, profit factor above 1.3, maximum drawdown below 10% and a sensible number of trades.
02 / Validate
Validate each strategy before authorizing execution.
03 / Authorize
The system is live. Validation and explicit activation remain requirements for execution.
Thresholds are validation criteria, not achieved results or a return guarantee.
Case study · Trading agent
Code has the final say.
Approve
A compliant proposal passes with an approved size.
Shrink
An oversized position is reduced to the applicable cap.
Block
A reached loss limit blocks the trade.
More than 1,100 automated tests, full trading-code coverage and independent adversarial review of risk, sizing and execution layers, per the technical brief.
Case study · Trading agent
Three ideas. Three risk decisions.
Illustrative scenarios. They show system behaviour, not financial results.
Checks pass. Code approves a sized position within the per-position cap.
Case study · Trading agent
Limits that stay in place.
No withdrawals
The agent never touches withdrawals or uses the client’s login.
Scoped keys
Keys are scoped to the intended trading permissions.
Explicit activation
Live mode requires a deliberate switch.
Approved sizes
Execution receives only risk-approved position sizes.
Loss limits
Per-position caps and the daily loss limit constrain execution.
Jurisdiction awareness
The system distinguishes US and Canadian rules.
Case study · Trading agent
Your business could be the next case.
Let’s explore a system built around your team’s work.
Lusona Autonomous, Montréal