Law Misi is our Legal Intelligence System. Enterprise-grade, closed-environment AI architecture.

It doesn’t guess.

• When a source is weak, it stays silent — and clearly shows what evidence is missing.

• It doesn’t repeat. Every answer is context-aware, tailored to the specific question and document environment.

• It’s never uncertain. Each statement includes a confidence score — a measurable indicator of reliability.

• Every answer is evidence-based. Cited with a legal reference point, paragraph ID, and date.

• All sources are traceable. An audit chain logs exactly why and how a conclusion was made.

• It understands relationships. Through a legal knowledge graph (a relationship-based learning model), it detects conflicts, omissions, and dependencies.

• When two rules collide, it highlights the contradiction and recommends the most reliable, up-to-date interpretation.

• It operates inside enterprise systems. Fully integrated with DMS / ERP / CRM platforms, respecting roles and permissions.

• It delivers only what matters for decisions — concise, verifiable, and actionable.

Team members:

1. Law Misi AI

2. Dr. Judit Lenke Tóth

3. Loránd Donkó

The Challenge

Law Misi’s task was to learn and interpret the complete set of local tax rules for all 3 183 municipalities in Hungary.

The source data came from the Hungarian State Treasury’s public system, where each municipality maintains its own Excel file — with different structures, field names, and logic.

Because downloads were only possible manually, an automated pipeline had to be built to fetch, interpret, and standardize data from 3 183 distinct sources.

The real difficulty: every municipality labeled the same concept differently — “building tax,” “local tax type,” “property tax” — meaning identical legal categories appeared under inconsistent terminology.

Law Misi therefore had to harmonize data structures, learning that different expressions and formats could describe the same legal concepts.

From this, Law Misi built a relationship-based knowledge graph — not merely storing data, but connecting it by legal context.

So when someone asks:

“Compare building taxes across Zala County in 2024,”

Law Misi not only answers precisely, but also:

• shows the evidence, 

• interprets the legal framework, and 

• compares practices among similar municipalities.

This development made it possible — for the first time in Hungary — for an AI system to see the entire municipal tax landscape clearly, transparently, and with genuine decision-support capability.