AI trading is not one technology
Products marketed as AI trading can range from simple rule-based automation to statistical models and machine-learning systems. Some generate signals for a human trader. Others place and manage trades automatically.
What matters more than the AI label
Start with measurable evidence: live-account history, length of track record, drawdown, number of trades, position sizing, profit factor, fees and where the trading account is held. A sophisticated description is not a substitute for a verifiable record.
Automation changes behaviour, not market risk
Automation can remove hesitation and emotional trade selection, but it cannot remove adverse market moves, execution risk or model failure. A system can follow its rules perfectly and still lose money.
How to evaluate an AI strategy
Separate the strategy from the marketing. Ask what instrument is traded, whether results are live or backtested, how much leverage is used, what the worst observed drawdown was and whether withdrawals and custody can be independently understood.