Money is both the goal and the instrument of illicit financial flows, terrorism financing and systemic financial attacks. Defending against them therefore has to be a financial discipline — and a scientifically grounded one. Silbad, Lile's financial-supervision product line, meets that bar with AI-first defence platforms for the institutions whose mandate it is to protect the European financial system.
Four fronts, six tools, one engine. A strict ethical and regulatory frame. Built in Austria, designed for the EU and the European Economic Area.
Silbad is built by Lile's interdisciplinary team of theoretical and applied scientists — physicists, mathematicians, philosophers, psychologists, computer scientists, economists and sociologists — who saw the same pattern from different angles: every illicit financial flow leaves a financial trace, and a sufficiently rigorous defence has to read those traces in real time, at scale, across actors, and with explanation good enough to stand up in front of a supervisor or a court.
Lindwurm reasons; Incitatus builds the models. Incitatus is Silbad's model-development infrastructure, built on Lindwurm: it continuously develops, evaluates and refreshes the domain models every Silbad product runs on — while detection and reasoning themselves are performed by Lindwurm, Lile's neuro-symbolic reasoning engine.
Banking, sovereign-market defence, listed-corporate cash-flow integrity and law-enforcement investigation — each is served by a dedicated product (Bucephalus, Marengo, Babieca, Chetak) tuned to its audience and regulatory environment, but powered by the same reasoning engine, Lindwurm.
Austrian rule of law, GDPR-by-design, banking and business secrecy respected by architecture not by promise, and an explicit ethical-distribution policy: we do not sell into countries whose human-rights standing is unstable. Every engagement is preceded by an ethical review.
Each line below is the entire product in one sentence. The deeper cards and the product pages take it from there.
The portfolio is split by audience because the regulatory frame, the language and the buying cycle differ for each — but the underlying reasoning engine, Lindwurm, and the ethical frame are the same.
Lindwurm reasons; Incitatus builds the models. Incitatus is Silbad's model-development infrastructure, built on Lindwurm: it continuously develops, evaluates and refreshes the domain models every Silbad product runs on — while detection and reasoning themselves are performed by Lindwurm, Lile's neuro-symbolic reasoning engine.
Explore Incitatus →Bucephalus delivers end-to-end IFF, AML and fraud defence for retail, commercial, investment and digital banks plus payment institutions across the EU and EEA. A federated architecture keeps customer data on-premises while the modelling improves continuously, and an ISO 20022 interface plugs into core-banking and payment systems with minimal disruption.
Explore Bucephalus →Marengo provides real-time defence against Systemic Financial Manipulation & Attack — a Silbad-defined category covering coordinated activity that targets a country's, an institution's or a company's stability, liquidity or market functioning. Marengo recognises both the manipulation itself and its preparation phase, and is offered exclusively to public-sector and EU institutional users.
Explore Marengo →Babieca will provide continuous cash-flow integrity monitoring for listed and pre-IPO non-financial corporates — operating in an industry-agnostic way, with a clear focus on the patterns of insider-assisted fraud, embezzlement and money laundering that traditional segregation-of-duty controls cannot reliably catch.
Explore Babieca →Chetak will support EU/EEA law-enforcement and prosecution authorities — police, the European Public Prosecutor's Office, OLAF, anti-corruption agencies, tax and customs investigation branches — with forensic-grade analysis of large offline datasets. The objective is to bend the asymmetry between modern data volumes and the human capacity available to read them.
Explore Chetak →Xanthos is a professional pseudo-transaction generator; Balius is its monitor. Together they let market participants and partners objectively test fraud and AML systems — including ours. Access is by registration; we ask for an optional, modest contribution which is dedicated to financial-literacy programmes for children.
These commitments are not a marketing layer over the products. They are the architectural and contractual frame within which the products are built and sold.
Lile, the venture behind Silbad, is headquartered in a traditionally independent, sovereign and neutral state. We focus on the EU, the EEA and the markets in economic community with them — while staying conversant with the international AML, IFF and CFT regimes outside the EU.
We do not say we will look after your data. We say your data does not come to us. The federated model keeps customer information inside its institution; only mathematical updates traverse the boundary. GDPR, banking secrecy and business secrecy are respected by design.
Every Silbad product produces a signal, a priority, a rationale and an evidence chain. None of them takes an operational decision on the customer's behalf. Aligned with the EU AI Act's human-oversight requirement (Article 14) and with GDPR Article 22.
We do not sell to companies or institutions in countries whose human-rights standing is unstable, where our products could enable abuse. Every sale and every knowledge-transfer engagement is preceded by an ethical review.
Four ways into Silbad. Pick the one that matches your reason for being here.
Silbad's detection and reasoning are powered by Lindwurm — Lile's neuro-symbolic reasoning engine. It learns from a handful of examples and produces decisions that can be explained, measured and independently audited.
Few examples. Traceable reasoning. Auditable outcomes.
In supervision, examples are scarce. Lindwurm is built for that world — it does not need oceans of data to reason well.
Every result traces back to data and method. Silbad surfaces a signal, a priority and an evidence chain — never a black box.
Decisions are built to stand up to a supervisor or a court, and to independent verification after the fact.
