Modern economic-crime investigations arrive with terabytes of seized data, multi-actor counterparty graphs, foreign-currency flows, encrypted comms metadata and statutory deadlines that do not move. The team facing it is small, expert, and human. Chetak is the forensic-grade layer that lets that team work at the speed of the case rather than the speed of the spreadsheet.
Chetak surfaces leads, evidence chains and explanations. The procedural decisions, the legal qualification, the charge — all of those remain with the investigator and the prosecutor.
Every modern economic-crime investigator lives between two curves. Without the right tooling, they trend against the case. Chetak's purpose is to bend them back.
As the seized data grows, the investigator's ability to hold the picture in their head falls in proportion — and error probability and analyst frustration rise to meet it. Chetak bridges that gap by reading the volume on the human's behalf and surfacing the items the human actually needs to see.
Every day the case stays open, the chance of proving it falls — and the chance that the offender repeats or that the evidence is destroyed rises in step. Chetak accelerates the discovery of the evidence chain so that time, for once, is on the investigator's side.
Three operational characteristics that define a Chetak deployment.
Banking statements, accounting books, e-mail metadata, real-estate registries, beneficial-ownership data and seized device extracts — Chetak ingests them together and surfaces the relationships across them. The investigator works with one connected picture, not seven disconnected exports.
Every Chetak finding arrives with the reasoning, the supporting evidence and the model-version stamp. The output is built to stand up in court — the opposite of the unaccountable "the algorithm said so" that defendants and appellate courts rightly reject.
Procedural, legal and charging decisions remain entirely with the investigator, the prosecutor and the court. Chetak's role is to make sure those decisions are taken on the basis of the evidence the case actually contains — not on the subset of evidence the team had time to read.
Chetak is offered exclusively to public-sector investigative and prosecutorial bodies operating in the EU, the EEA and the markets in economic community with them — and only after a Silbad ethical review.
The country-level investigative function for fraud, money laundering, corruption, tax-crime proceeds and organised economic crime — the largest single audience for the platform.
The EU-level prosecutorial body for crimes against the financial interests of the Union. Chetak is built to handle the scale and the cross-border structure of EPPO's typical caseload.
The EU's investigative function for fraud, corruption and serious irregularities touching the EU budget and its institutions.
Specialised anti-corruption authorities investigating public-sector and high-value private-sector corruption.
Investigative arms of national tax authorities — for example, in Hungary the tax authority's investigation directorate (NAV PEI). Chetak is positioned for the criminal-investigation track, not for routine tax compliance.
Customs authorities' economic-protection arms investigating excise fraud, trade-based money laundering, dual-use export evasion and similar.
Chetak is not sold outside the EU/EEA, and not to authorities operating in jurisdictions whose human-rights and rule-of-law standing Silbad's ethical review judges unstable. The screening is procedural — it precedes every contract, including renewals.
Investigative-analytics platforms exist. Most of the largest are US-headquartered, built on analytics-first architectures that retrofit AI as a module, and operate under data-residency frameworks the EU's recent jurisprudence has not been kind to.
Chetak is built and operated by Lile, an Austrian venture under EU law. There is no non-EU parent jurisdiction with a Cloud Act-style claim on the data the system processes — and the federated architecture keeps customer data inside the customer institution to begin with.
Chetak runs on Lindwurm, Lile's reasoning engine; its domain models are built by Incitatus, Silbad's model-development infrastructure. Models are refreshed continuously, so investigators receive protection that is no older than the last refresh.
Investigative AI sits in the EU AI Act's high-risk category, with strict human-oversight, transparency and explainability obligations. Chetak's architecture treats those as the default, not as a configuration to be enabled — there is no "autonomous" mode to switch on by accident.
Every Chetak finding can be traced to its inputs, its model version and its reasoning — the form of explainability courts and appellate review require. Findings without that trail are flagged as suggestive, never as evidence.
Chetak is the retrospective, forensic-mode product in the portfolio. The other three deployable systems work in real time, on different audiences, but share the same reasoning engine, Lindwurm, and the same model factory, Incitatus.
Chetak engagements are bespoke. The first conversation is structured discovery: the scale and structure of the data the institution typically faces, the evidentiary standards the prosecutor's office requires, the human-oversight model the institution is bound by, and a realistic pilot envelope.
