Summary
FinHarbor, a technical platform provider for compliant financial products, has launched an AI Act-ready compliance module — an AI co-investigator that works on top of the platform's existing compliance stack and is deployed entirely inside the client's own infrastructure. The self-hosted LLM connects to the platform's ledger, KYC/KYB, KYT, transaction monitoring, and append-only audit trail.
The module addresses a well-documented problem: more than 95% of alerts from rules-based AML systems are false positives, and roughly 98% never result in a suspicious activity report. The AI co-investigator pulls client and transaction data on demand via plain language queries, clears routine false positives with documented reasoning, assembles full investigation cases, drafts SAR/STR narratives, and prioritizes the investigation queue by risk score. Crucially, it never files a SAR, blocks an account, or takes any action with regulatory consequences on its own — every such step requires a human signature.
Key Facts
- Module: AI co-investigator for AML compliance
- Architecture: self-hosted LLM inside client's environment
- Connected to: ledger, KYC/KYB, KYT, transaction monitoring, audit trail
- Problem: 95%+ false positive rate in rules-based AML systems
- Capabilities: plain language queries, false positive clearance, case assembly, SAR drafting
- Human oversight: AI investigates, humans decide; no regulatory action without human signature
- EU AI Act: designed around the Act's logic (documented, supervised, incapable of acting alone)
- DORA compliance: append-only audit log meets third-party oversight requirements
- Data residency: regulated data never leaves client's perimeter
- Jurisdictions: covers GDPR, UK GDPR, revFADP, LGPD, PDPL
- CEO Ilya Podoynitsyn: "Compliance teams don't need another dashboard — they need the routine taken off their hands without giving up control"
- Status: running in pilots on several client projects
Why It Matters
FinHarbor's AI co-investigator addresses one of the most expensive operational problems in financial services: the overwhelming volume of false positive AML alerts that consume compliance team resources. By taking over the routine layer while keeping humans in the loop for decisions with regulatory consequences, the module offers a practical path to AI adoption in compliance — one that satisfies both efficiency goals and regulatory requirements like the EU AI Act and DORA.