Fewer false positives
Models learn what normal looks like for each group, so legitimate active traders aren't flagged alongside abusers.
RNIX AI Desk combines proven rule-based detection with machine-learning models trained on trading behaviour, and explains each decision so your desk can act with confidence.
Orders, deals, positions, logins, IPs and prices stream in from MT5 or cTrader through read-only APIs.
Pattern detectors for HFT, latency, martingale, collusion and more evaluate each event instantly.
A model weighs all signals against each account's history and your book to produce one 0–100 score.
Plain-language reasons, suggested A/B-book routing and next actions go straight to your desk.
Each module produces its own score. The AI layer combines them using weights learned from your book: what normally leads to losses, which signals tend to appear together and how each group of clients usually behaves.
The composite score is continuous, from 0 to 100, and moves as the account trades. It is grouped into four bands so your team knows what to do at a glance.
Illustrative example.
Sample conversation for illustration.
The built-in copilot answers questions about your book in plain English. Ask which accounts drove today's losses, why a client was flagged or what exposure looks like ahead of a news release, and get an answer backed by the underlying data.
Models learn what normal looks like for each group, so legitimate active traders aren't flagged alongside abusers.
As abuse patterns change, models are re-tuned without you rebuilding rules by hand.
A/B-book suggestions come with the exact signals behind them, ready for dealing and compliance review.
AI recommends, your team decides. Every automated action can require approval and is fully logged.
Models run inside your deployment. Your trading data is never shared or used for other clients.
Weights, thresholds and windows can be set per group, account type or promotion.
No. Every score is shown with the signals that produced it and a plain-language explanation, so dealers and compliance can see exactly why an account was flagged.
It learns from your own platform's trading data: deals, orders, positions, logins and price feeds. Your data is not shared with other clients.
Only if you allow it. By default the AI recommends and your team approves. Automatic actions can be enabled per rule with full logging.
Models are re-tuned during onboarding with your risk team and updated continuously as new abuse patterns appear.
See RNIX AI Desk running against your own MT5 or cTrader data in a live walkthrough with our team.