Our Story
From Research Lab to Global Impact
The Beginning
Filuta AI was born from a simple observation: the world's most complex software systems were being tested with the simplest tools. Scripted tests, manual QA, and checkbox-driven compliance couldn't keep up with the pace of modern development — or the complexity of real-world interactions.
Our founders, a team of AI researchers and engineers with deep roots in symbolic AI planning, saw an opportunity to change that. Drawing on decades of combined experience at institutions and organizations working on NASA-grade planning, DARPA research, and cutting-edge AI conferences, they set out to build something fundamentally different: AI agents that could test software the way humans do — but at machine scale.
Neurosymbolic AI: Our Core
At the heart of Filuta is a neurosymbolic engine — Gartner's name for the category is Composite AI. Symbolic planning supplies precision, auditability and an oracle that is independent of the system under test; machine learning supplies perception and adaptation. Language models can read everything and prove nothing; symbolic AI can prove but cannot read the world. The engine sits between the two.
Every action is explained, every finding reproduces on demand, and every run ends with a coverage number and a statement of exactly what was proven — from a single feature to a full release train, without losing traceability.
Where We Are Today
Founded in 2022 and headquartered in Prague, with a US entity in Texas, Filuta AI operates at the intersection of research and production delivery. The team is research-heavy — AI scientists and engineers, several of them professors — and holds one of the largest concentrations of symbolic AI planning expertise anywhere.
The engine is in production for cockpit software at a European automotive group — nine years of manual tester-time now runs in ten days per release — and for procurement-compliance screening at Houston ISD in Texas. The same engine extends to games, simulations and the AI agents that will soon operate our interfaces.
What Drives Us
For all of history, “did you check everything?” had one honest answer: “I did my best.” For regulated software that answer is no longer accepted — evidence is required, per release. We exist so that whoever has to sign that an interactive system is safe to release can sign on proof rather than on hope, one hypothesis at a time.
One engine. Whoever signs that the release is safe.
With Less QA Overhead
