GovernmentOperations

How MPRS Uses NVIDIA AI to Resolve Legal Cases in Minutes Instead of Months

The Public Ministry of Rio Grande do Sul (MPRS) in Brazil, serving over 8 million citizens across 497 municipalities, partnered with WideLabs to deploy AI agents built on NVIDIA NIM and NeMo that process the state’s 60,000 annual police inquiries and 39,000 public service requests. Legal procedures that previously took months or years are now resolved in less than a day.

Outcomes

Months/years to less than 1 dayInvestigation resolution time
8 million+Citizens served
60,000Annual police inquiries processed
39,000+Annual public service requests

Tools & Technologies

1NN
NVIDIA NIM
NVIDIA Inference Microservices — containerized, optimized inference endpoints for deploying AI models at production scale.
2NN
NVIDIA NeMo Guardrails
Toolkit for adding programmable safety guardrails, topic filters, and dialog flows to LLM-based applications.
3NN
NVIDIA NeMo Retriever
NVIDIA NeMo Retriever provides embedding, reranking, and hybrid search capabilities for building production-ready RAG pipelines.
4NA
NVIDIA AI Enterprise
NVIDIA AI Enterprise is an end-to-end platform for deploying and managing AI workloads, including NeMo models, NIM microservices, and GPU-accelerated infrastructure.
5NN
NVIDIA NeMo Curator
GPU-accelerated data curation library for deduplication, filtering, and preprocessing LLM training datasets.

AI Categories

Challenge

MPRS processed roughly 60,000 police inquiries and 39,000 public service requests annually using manual, query-dependent systems that slowed investigations to months or years, while citizens—especially vulnerable populations—had no accessible interface to the justice system.

Solution

WideLabs built the Justice Intelligence platform on NVIDIA AI Enterprise, deploying three AI agents using NVIDIA NIM microservices, NeMo Retriever RAG pipelines, NeMo Curator for data preparation, and NeMo Guardrails for safety and consistency, with all models running on locally controlled sovereign infrastructure.

Full Story

Brazil’s justice system is under extraordinary strain. With over 82 million pending legal cases nationwide as of late 2023, and Rio Grande do Sul alone archiving roughly 60,000 police inquiries each year, the Public Ministry of Rio Grande do Sul (MPRS) faces a caseload that stretches the limits of conventional legal process. The MPRS also handles approximately 39,000 public service requests annually, a number that continues to grow as citizens increasingly exercise their legal rights. Investigators, prosecutors, and analysts were spending inordinate time on routine document searches and archiving tasks that demanded both legal expertise and database fluency.

Access 399+ AI use cases, 401+ tools, and adoption signal rankings.

Source

NVIDIA
January 2025
Original case study

Similar Cases

1C
How CACI's DarkBlue Uses Elasticsearch and Claude to Accelerate Dark Web Criminal Investigations
CACI
Seconds per query regardless of data age or volumeCriminal investigation acceleration
2NM
How New Mexico County Uses C3 AI to Boost Property Appraisal Accuracy by 50%
New Mexico County Assessor
+50 percentage pointsImprovement in model accuracy
3M
How ManTech Cut IT Tier 1 Support Workload by 50% with Moveworks AI
ManTech
50%IT Tier 1 FTE reduction
4ME
How Massachusetts Education Office Saves $1.5M Annually with Snowflake
Massachusetts Executive Office of Education
$1.5MAnnual cost savings from Oracle to Snowflake migration
5NH
How NYC Health + Hospitals Uses Snowflake to Cut Data Delivery from Days to Minutes
NYC Health + Hospitals
5 days → 5 minutesData delivery time for membership and claims data
6T
How ThinkDeep Used NVIDIA AI to Cut Government Document Processing from 2 Days to 2 Minutes
ThinkDeep
2 days → 2 minutesDocument processing time
7ME
How Massachusetts EOE Uses Snowflake to Save $1.5M Per Year
Massachusetts Executive Office of Education
$1.5 millionAnnual cost savings from Oracle migration
8R
How Recursion and MIT Built Boltz-2 to Cut Drug Discovery Wet Lab Work 40% with NVIDIA
Recursion
40%Wet lab work reduction for equivalent value
9SD
How Seattle DOT Uses C3 AI to Cut Collision Analysis Time by 90%
Seattle Department of Transportation (SDOT)
90%+Reduction in collision analysis time
10GC
How Garden City Public Schools Uses Samsara to Cut Maintenance Costs by 66%
Garden City Public Schools
$150,000+Annual cost savings
See all use cases →