Your compliance stack detects.It doesn’t investigate.
Novaheim builds AI that turns an escalated alert into a complete case file and a draft SAR, with fiat and on-chain activity in one record and every fact cited to its source. Your investigator reviews it. Your BSA Officer decides and files.
- Runs in your cloud
- SAR data never leaves it
- Your officer signs every filing
The alert is the cheap part.
Detection gets the budget. The real cost is the investigation that follows.
Open the case. It’s already built.
On March 2, Harlow Trading received a wire of $48,500[A1] and within four hours sent $48,200 in USDC to an external wallet[T4]. Funds reached a high-risk exchange in two hops[C2], inconsistent with the stated business purpose[K1].
Activity type · 31z
Schema · valid
AI drafts. Code checks. Your officer decides.
AssembleOne case fileAI
Meet the founder
I build the systems that handle what happens after the alert fires.
My background is statistics and applied analytics: seven years measuring model performance, false positive rates and where automated systems fail. That discipline is what AML programs need and rarely have. Monitoring stacks get tuned on intuition, and few teams can show an examiner why a threshold sits where it does.
Read more
I work across BSA/AML, OFAC, the FATF Travel Rule and the GENIUS Act stablecoin rules, and I read chains directly rather than relying on a vendor’s attribution. Most compliance stacks see the fiat leg and go blind at the chain boundary. That gap is why Novaheim exists.
Questions a BSA Officer asks first.
No. It drafts and assembles. Your investigator recommends, and your BSA Officer decides and files, exactly as today.
Everything runs in your cloud, including logs and test results. Nothing is copied to us or used to train models, access is limited to named roles, and we agree the controls in writing before we see a case.
Your BSA Officer, through your own BSA E-Filing account. We never hold your filing credentials.
Each vendor sees its own data. We join the fiat and on-chain sides into one cited case, on top of the tools you already pay for.
We expect a few hours of your team’s time across the build, mostly in the diagnostic and backtest grading.
The diagnostic shows which wins are realistic for your program, measured on your own cases.
Backtest results, citation checks and known limits are documented in a validation pack your model risk review can use and hand to an examiner.
You do. The system runs in your environment and stays there.
Alert triage, which often comes first because it feeds the case. Also KYB periodic review, and a policy agent grounded in your own program documents.
The backlog isn’t an efficiency problem. It’s a finding.
When alert volume outruns team capacity, the 30-day clock doesn’t move. Let’s find the workflow costing you the most and build the system that clears it — with an audit trail on every determination.