The line your book lives by — watched in real time.
Sentinel is automated oversight: continuous exposure monitoring, proprietary quant/ML detection of toxic flow and abuse, and workflows that act when a signal crosses the limit — before small signals become expensive. Included in Vanquor CRM.
Risk is not one number
Surveillance runs across exposure, behaviour and flow quality simultaneously — on the same live data your platforms produce.
Real-time risk monitoring
Live exposure across books, symbols, groups and accounts — drawdown, concentration, correlation, leverage — evaluated continuously against configurable limits. Breaches trigger workflows in seconds, not at end-of-day reconciliation.
Toxic-flow classification
Proprietary models score flow as it develops: latency and stale-quote arbitrage, news-window exploitation, gap and rollover strategies, coordinated behaviour, evaluation gaming. Classification feeds routing, review and enforcement — with the evidence attached.
Abuse & anomaly detection
Multi-accounting signals, opposite-side hedging across entities, bonus abuse, and statistical anomalies that don’t match a named pattern yet — built by the quant-research team for trading operations, not generic fraud heuristics.
Case management & audit
Every alert becomes a case with its evidence chain: what fired, what the models saw, who reviewed it, what action was taken. The audit trail is a by-product of the workflow, not an extra job.
From signal to action
Surveillance that only reports is a dashboard. Sentinel closes the loop.
- Your limits, your policy. Thresholds, severities and response playbooks are configuration, not code changes.
- Honest automation. Fully automated where you want speed; human-in-the-loop where judgement matters — always recorded.
- AI investigation. Claude and GPT (via MCP) explain what fired and why, build cross-account timelines, and answer plain-language questions about the flow record.
See your own flow the way Sentinel sees it.
Bring a real scenario to the demo — an abuse pattern you've fought, a breach that got expensive — and we'll walk through how detection and workflow would have handled it.