Our mission
Our mission is to ensure that people in the Czech Republic have practical access to accurate, verified information about the legal, financial, and bureaucratic systems that govern their lives.
If systems like Mendel are built responsibly and deployed at scale, this technology could help give everyone access to information that currently requires either professional expertise or hours spent reading documents written for someone else. The practical benefits are not abstract: better-informed decisions, reduced dependence on expensive intermediaries, and a meaningful increase in what ordinary people can ask of the institutions around them.
We can imagine a world in which a person facing a landlord dispute or a tax question has access to something that knows the relevant law and can explain it clearly, without inventing a detail that sends them in the wrong direction. That kind of access does not exist today. We believe it should.
On the other hand, a system that presents incorrect information with authority causes real harm. In legal and financial contexts, that harm is often invisible: wrong decisions made in good faith, obligations missed, professionals given a flawed foundation to work from. This risk shapes every architectural decision we make.
Because the potential benefit is significant, we do not believe it is possible or desirable to avoid building this technology. The question is not whether systems like this will exist. The question is whether they will be built with the rigor the domain demands.
Although we cannot predict exactly how these systems will develop, we can state the principles we hold most consistently:
- We want practical access to information about Czech law, Czech finance, and Czech public administration to be available to every person in the country, not just those with professional training or the resources to pay for it.
- We want every system we build to be honest about the limits of what it knows. When it reaches a question it cannot answer reliably, it should say so, rather than fill the gap with a plausible-sounding response.
- We want to develop these capabilities in a way that strengthens the professional institutions that carry real accountability, rather than degrading the relationships that make those institutions function.
The near term
There are several things we think are important to do now.
First, we are building Mendel, our initial product for the Czech market, around a single architectural constraint: every answer must trace to a verifiable source. The system exposes its reasoning and marks uncertainty explicitly. We believe this is the right foundation because earning trust in a domain this consequential requires accountability to be built into the architecture from the start. We plan to deploy it, learn from real use, and iterate.
Second, we are investing in research on autonomous self-verification: systems capable of checking their own reasoning before presenting a conclusion, surfacing that reasoning transparently, and flagging where uncertainty is highest. Our goal is a system that knows when it is likely to be wrong and tells you so, not a system that never appears uncertain. In legal contexts, those two targets require very different choices.
Third, we think the question of professional displacement deserves a direct answer. If a machine can answer the questions people currently pay lawyers and financial advisors to answer, some of those professionals will lose work. We take that seriously, and our response is structural. The most capable features of the system, those approaching professional-level synthesis rather than information access, are available only to verified, licensed practitioners. A lawyer who uses Mendel to process case law before a hearing becomes a better lawyer. A citizen who uses it to understand their rights becomes a better client. The judgment, the accountability, and the professional relationship stay with the human. The tool amplifies what that human brings to the work.
We believe that access to information should function as a right, not a premium service. We also believe that professional judgment is not something software should carry. The product is designed to hold both of those positions.
The long term
We believe that the systems we are building are the beginning of something much larger.
The legal and bureaucratic systems of Central Europe share common roots. Czech and Slovak law developed from the same foundation. The EU regulatory framework applies across the continent. The problem we are solving in Prague exists in Warsaw, Bratislava, Vienna, and Bucharest in structurally identical form. We expect the technology we are building to serve those markets, and we are building with that scope in mind.
Our longer investment is in language models developed specifically for Central and Eastern European legal and administrative traditions. General-purpose models are capable. They were not trained in depth on the texts and interpretive traditions of this part of the world, and the gap shows in precision exactly where it matters most. A model built for Czech legal text and Slovak administrative procedure, trained on the corpus that actually governs people’s lives in this region, will outperform a general system on these tasks in ways that translate directly into outcomes for the people using it.
We think it is important that this work be done by people who understand the context, operate under European law, and have made the architectural commitments to accuracy and transparency that the domain requires. We intend to be those people.
The information people need to make sense of their own legal and financial situations is, in most countries, technically public. We think making it practically public, for everyone, is worth building a company around.
CAIROS Czech Republic hello@cairos.cz
