Since 2018 Brazil's Chamber of Deputies has run Ulysses, an eight-module AI suite matching similar bills (20,000+ requests/yr), transcribing speeches, analysing comments and authenticating votes biometrically; a 2025-26 relaunch added AI governance rules.
20,000+ requests/year
U3 bill-similarity requests processed
Details
Maturity
Established
Promoter
Câmara dos Deputados (Brazilian Chamber of Deputies)
Period
2018–present
Keywords
legislative technology, public administration, natural language processing
Context
Brazil's Chamber of Deputies has operated Ulysses, an in-house suite of AI and NLP tools for legislative work, since 2018, growing out of the Chamber's Digital Strategy (Act 268/2021) into eight modules (U1-U8).
Activities
U3, the most heavily used module, runs semantic analysis to flag bills similar to a new proposal; other modules cluster amendments, transcribe plenary speech, analyse public e-poll comments, and apply biometric facial recognition to authenticate votes and tag legislators in photographs. In December 2025 the Chamber launched an internal 'Ulysses Chat' assistant and new AI governance and security policies.
Results
U3 processes more than 20,000 bill-similarity requests every year.
Implementation
Indicative cost
Medium (€50k–€500k) — In-house suite of 8 modules developed since 2018 under the Chamber's Digital Strategy (Act 268/2021); relies partly on Microsoft Azure and a planned OpenAI Whisper integration.
Time to results
Long (> 3 years) — Operating since 2018; new phase with governance policies and internal chat assistant announced December 2025.
Staffing & skills
Chamber of Deputies Digital Strategy team, AI governance and security policy staff (2025-26 relaunch)
Conditions for success
in-house NLP/AI development capacity
formal governance policy for external AI tool use by staff
Common failure modes
biometric voting and facial-recognition modules (U7/U8) operate without a named external oversight body
no independent accuracy evaluation published
Commonly funded by
National / regional programmes
Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.
Do you run this practice?
Claim it —
verified implementers get a public contact pathway and can propose corrections.
Data sources
Where this practice's information was retrieved from, and when.
Baden-Württemberg built F13, the first AI assistant in productive use across a full German state administration, running on sovereign infrastructure; …
