About

Human cognition is remarkable. Language is its most visible trace — in texts, speeches, narratives, and the metaphors societies use to make sense of the world. At Crow Intelligence, we study that trace. We use computational methods and cognitive science to understand how language shapes thought, how narratives drive decisions, and how technology is transforming both. We believe in open inquiry, rigorous method, and the Popperian ideal of an open society. We think technology is not neutral — it mediates and shapes social cognition in ways that are worth studying seriously and honestly. Ἡ σοφία ἄρχεται ἀπὸ τῆς ἀπορίας. — Wisdom begins in wonder. (Socrates)

What We Do

We conduct original research at the intersection of cognitive linguistics, natural language processing, and data analysis. Our work combines academic depth with computational tools to reveal patterns in how people and institutions think, communicate, and decide.

We publish our findings openly. We also take on research and analysis commissions for organizations that need this kind of expertise — specification, evaluation, and interpretation of language-based AI systems.

We do not build software. We help you think clearly about what to build, whether it is working, and what it means.

Orsolya Putz, PhD

Orsolya Putz, PhD

Orsolya is a cognitive linguist and AI engineer. She holds a PhD from Eötvös Loránd University and is an adjunct professor at the Budapest University of Technology and Economics (BME).

Her book Metaphor and National Identity: Alternative Conceptualization of the Treaty of Trianon (John Benjamins, 2019) examines how conceptual metaphors structure collective memory and national identity in Hungarian political discourse — tracing the same underlying metaphors surviving across a century of public speech. She has published in cognitive linguistics, cultural linguistics, and multimodal metaphor research.

Her research asks a deceptively simple question: when societies reach for the same metaphors across generations, what does that reveal about the structure of collective thought?

Zoltán Varjú

Zoltán Varjú

Zoltán is an NLP engineer, data scientist, and serial entrepreneur with twenty years of experience building language-based systems.

He co-founded and led two projects funded by the Google Digital News Initiative:

  • Source Code Leak — a digital fingerprinting tool for investigative journalists, revealing hidden connections between seemingly unrelated organisations. It evolved into Complytron, an AML compliance platform acquired by SEON in 2023 (covered by TechCrunch, Finovate, and PYMNTS).
  • Koffair / Radioship — open-source speech-to-text infrastructure for community radio stations, making large audio archives searchable for the first time (GitHub). Built with Tilos Rádió and Civil Rádió.

He has been a long-time author and data collaborator at Nyelv és Tudomány (nyest.hu), Hungary's leading science communication platform for linguistics. He has worked across fintech, regtech, media analytics, and green finance — always at the intersection of language, data, and decisions.

Why Crow Intelligence

We started Crow Intelligence because we kept finding ourselves drawn to the same kind of question: what does language reveal about how people and institutions actually think — beneath the official version?

The tools of modern NLP make it possible to study that question at scale, across centuries of text, across thousands of speeches, across the full arc of a financial crisis as it unfolds in the news. Cognitive linguistics gives us the theoretical framework to know what we are looking at when the data speaks.

That combination — computational power grounded in serious linguistic and cognitive theory — is still rare. That is the gap we are trying to fill.

Follow the Research