Live IBM synthetic AML dataset · 1M+ transactions · Full audit trail

$3.1 Trillion
in Illicit Funds.
One Deterministic
Evidence Trail.

The Axiomyx Platform builds the evidence ledger before AI writes a word — canonical payload, β/Df/Dh signature, Zone v2 placement, graph path, and reproducibility hash. The AI narrative is bound to those deterministic outputs. No governance pressure. No contested evidence. No exceptions.

$3.1T
Illicit funds flowed through the global financial system in 2023 (Nasdaq Verafin)
~1%
Of laundered funds ever detected or seized
100%
Axiomyx audit trail coverage — every alert, every decision
5
Deterministic platform modules — Canonical Pipeline to Graph Provenance
The problem today

Current AML AI fails banks in three ways

Many AML AI workflows depend on probabilistic scoring, embeddings, or free-form narrative generation. That creates governance pressure: investigators need to know why a transaction was flagged, what evidence was linked, and whether the result can be reproduced.

🎭

Hallucinations

AI invents account relationships that don't exist. A layering scheme across five accounts gets misread as three unrelated transactions. The SAR narrative is wrong. The regulator rejects it.

🌀

Drift & False Positives

The same $9,847 transfer scores HIGH RISK on Monday and MEDIUM RISK on Friday. No prompt engineering system can guarantee repeatable outputs. Compliance teams stop trusting the alerts.

🕳️

No Audit Trail

The regulator asks: "Why did your system flag this transaction?" The answer is: "The model thought so." That is not an answer. Regulators require deterministic, reproducible reasoning. Every time.

The Axiomyx solution

Five modules. One deterministic money trail.

Click any transaction on the left and watch the full Axiomyx Platform process it in real time — from raw record to evidence-bound AI investigation narrative. No guessing at any layer.

IBM AML Dataset — Live Sample
Processing

Dataset: IBM public synthetic AML transaction data (IBM/AML-Data). CSV financial transactions with laundering tags used for benchmark evaluation. Repository licence: Apache-2.0. Data licence: CDLA-Sharing-1.0. Independent Axiomyx benchmark — no IBM endorsement implied.

Live platform output

What the Axiomyx Platform produces — in real time

Every transaction processed through the 5-module pipeline. Geometric zone placement. Cross-account entity matching. Alert feed with evidence-bound AI investigation narratives. All deterministic.

📊
Platform Metrics
Live processing stats — IBM dataset
25
Processed
24
Flagged
24
For Review
7B-S2
Hot Zone
100%
Reproducibility
12ms
Avg Pipeline
Geometry Core — Zone v2 Map (Dh × β)

🔴 Suspicious cluster   🟡 Under review   🔵 Normal pattern

🚨
Live Alert Feed
AI-written SAR-style investigation narratives · Full audit trail
Full audit trail

What the regulator sees. Every time.

Every transaction. Every module output. Every AI narrative. Timestamped. Deterministic. The same input always produces the same output. That is what auditable means.

Timestamp Transaction ID Amount β / Df / Dh Zone Entity Match Risk SAR Narrative
Next step

The dataset is synthetic.
The platform is real.
The audit trail is yours.

Everything you just saw runs on the Axiomyx Platform — five deterministic modules processing transactions into fully auditable, AI-native intelligence. The evidence ledger is built before the AI writes a single word.

Request a platform briefing View trading agents demo

Also live: EUR/USD and GBP/JPY paper trading agents proving the same platform across financial markets. View live →