PRODUCT DESIGN PERSPECTIVE

Leila Haddad

AI & Quant Safety

The question this perspective asks

Do AI and automated strategies stay inside the limits the user chose?

This perspective treats AI and strategies as limited tools. Their data, actions, accounts, budgets, and risk limits should be explicit, testable, and reviewable after a run.

What it checks

Allowed toolsThe agent can call only approved tools.

Test before liveStrategies run in backtest or simulation first.

Reviewable recordInputs, decisions, actions, and limits are logged.

Back to About us Read the user guide