Use Case Intelligence RSS Search use cases Search
NVIDIA Public Sector Brazil

Public Ministry of Rio Grande do Sul (MPRS)

Document Processing · Enterprise Knowledge Search · Security and Compliance

MPRS needed to manage rising legal caseloads, archived police inquiries, and public-service requests more efficiently, but the volume of legal documents and the need for compliant, locally controlled analysis made traditional manual and standard RAG workflows too slow and unreliable.

Read the original story Explore all use cases
497 municipalities
About 60,000 police inquiries archived each year in Rio Grande do Sul
Around 39,000 public service requests annually
More than 8 million citizens

Reported impact

Business outcome

The platform moved investigations that could last months or years to less than a day, improved transparency and access for more than 8 million citizens across 497 municipalities, and created a compliant AI foundation for corruption analysis, case archiving, and public legal assistance.

  • Compliance
  • Cost Reduction
  • Customer Experience
  • Productivity Improvement

Solution

WideLabs built the Justice Intelligence platform for MPRS on NVIDIA's NeMo stack to support investigations and citizen-facing legal access. TORI Investigation Assistant uses NVIDIA NeMo Retriever for embedding, reranking, and hybrid search inside retrieval-augmented generation pipelines deployed as NVIDIA NIM microservices against legal databases. Archiving Intelligence Assistant applies natural-language analysis to police inquiries and archiving decisions, while the Citizen Access Agent uses NeMo Guardrails for secure routing, case-type determination, document management, and appointment scheduling. Under the hood, NVIDIA AI Enterprise provides the managed software foundation, NeMo Curator preprocesses tens of thousands of documents for retrieval, and NeMo Customizer plus NeMo Aligner adapt models to the local legal domain. The result is a locally controlled, compliance-oriented legal AI platform that combines RAG, guardrails, and curated data pipelines for prosecutors and the public.

Data flow

Legal records, police inquiries, and citizen requests enter the Justice Intelligence platform, NeMo Curator prepares the document base, NeMo Retriever and NIM drive RAG search and agent responses, Guardrails governs routing and safe interaction, and the results are returned to prosecutors, analysts, or citizens through investigation and access assistants.

Solution architecture

11 components · 4 layers
  1. Compute
    • NVIDIA NeMo Retriever Powers embedding, reranking, hybrid search, and document chunking for legal retrieval.
    • NVIDIA NeMo Curator Preprocesses large and growing legal-document collections for high-quality retrieval.
  2. Serving
    • Justice Intelligence platform Acts as the main legal AI environment used by MPRS for investigations and public-service workflows.
    • TORI Investigation Assistant Provides prosecutors and analysts with natural-language access to legal data and investigation records.
    • Archiving Intelligence Assistant (AIA) Analyzes police inquiry archiving decisions and explains them in accessible language.
    • Citizen Access Agent Guides citizens through secure legal-service access, routing, and scheduling workflows.
    • NVIDIA NIM Deploys the RAG and agent models efficiently as scalable microservices.
  3. Orchestration
    • NVIDIA AI Enterprise Provides the locally controlled software foundation needed to manage the legal AI stack.
    • NVIDIA NeMo Customizer Adapts models to the MPRS legal domain and local requirements.
  4. Governance
    • NVIDIA NeMo Guardrails Enforces safe and deterministic behavior for citizen and investigator interactions.
    • NVIDIA NeMo Aligner Improves domain understanding for specialized legal and auditing terminology.

Architecture clues

  • Local control over data and models was a design requirement due to legal and regulatory constraints.
  • MPRS runs separate agent experiences for investigators and citizens on top of the same NeMo-based foundation.
  • The platform uses enhanced RAG with hybrid search, reranking, advanced chunking, and legal-domain model customization.

Evidence from the source

"Procedures and investigations that could last months or years can be resolved or advanced in less than a day."
A scalable, GPU-accelerated pipeline for document preprocessing was developed with NVIDIA NeMo Curator.
NVIDIA NeMo Guardrails work to ensure consistent information retrieval and behavior.
Natural language query capability for complex database searches, achieved using advanced retrieval-augmented generation (RAG) pipelines built with NVIDIA NeMo™ Retriever ... and deployed efficiently as NVIDIA NIM™ microservices.
The Justice Intelligence platform developed by WideLabs for MPRS consists of two sophisticated AI agents designed to revolutionize how prosecutors, attorneys, analysts, and assistants at MPRS handle cases and investigations.
Transparency Evidence & methodology 100/100 reference value · 89% confidence · 1 source

Review how this reference was scored and trace each normalized claim back to its documented source.

01

Scoring methodology

How scores are calculated

Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.

Customer reference value100/100
  • Problem clarity20/20
    Measures whether the customer's operational problem is specific enough to compare.
  • Implementation depth25/25
    Rewards named components, data flow, integrations, and architecture detail.
  • Outcome specificity20/20
    Measures whether the reported benefit is concrete rather than a generic claim.
  • Quantified impact25/25
    2 metrics describe measurable impact.
  • Deployment scale10/10
    Rewards evidence of a live or scaled deployment and meaningful integration scope.
Evidence confidence89%
  • Source authority20/25
    The main evidence is an official vendor-published customer story.
  • Claim traceability21/25
    10 of 12 key fields link to a named source.
  • Direct evidence20/20
    5 direct source quotes retained.
  • Date provenance10/10
    The publication date comes from the source page or publisher.
  • Source transparency8/10
    1 source reference document publisher, URL, authority, retrieval, and supported fields.
  • Implementation specificity10/10
    Rewards source-backed components, data flow, quantified impact, and a specific outcome.
02

Traceability

Sources & provenance

Field coverage shows which facts each source supports. Dates are shown only when documented by the source.

  1. NVIDIA

    Public Ministry of Rio Grande do Sul (MPRS) customer story

    First-party
    https://www.nvidia.com/en-us/case-studies/widelabs-ai-makes-legal-services-accessible-brazil/
    Source type
    Vendor-published customer story
    Published
    Jul 7, 2026
    Supports
    • Customer
    • Business problem
    • Solution
    • Products
    • Solution components
    • Data flow
    • Integrations
    • technical area
    • use case category
    • Business outcome
    • outcome category
    • Metrics
    • Architecture
    • evidence quotes
    • Publication date

Date evidence

  • Published on source page Jul 7, 2026 · NVIDIA · First-party
    Open date source
Not separately sourced Industry, Region, Company size. Treat these normalized attributes as lower-confidence until a dedicated source is added.

Related

Similar use cases

See all like this
AWS Apr 15, 2024 InfoCert Public Sector Italy Training and deploying these models was a lengthy, costly process, which limited access to this set of capabilities to larger companies. While InfoCert was searching for a way to streamline such ML integrations, AWS… Additionally, InfoCert expects to reduce financial overhead by 80 percent for startup, development, and… Document ProcessingEnterprise Knowledge Search Amazon Bedrock NVIDIA Jul 7, 2026 French Ministry of Economy and Finance Public Sector France French public servants needed to answer citizen requests against more than 100 million documents per year while maintaining legal accuracy, security, confidentiality, and compliance, but manual retrieval and analysis of… 1,000 work hours saved per month for a typical 100-user deployment Document ProcessingFraud Detection DeepBrain · NVIDIA AI Enterprise · NVIDIA DGX platform NVIDIA Jan 17, 2025 National Informatics Centre Public Sector India NIC needed to reduce judicial backlog and improve accessibility in a court system with massive case volumes and many official languages, while giving judges faster ways to search and summarize legal material. 18,000 daily district court sessions Document ProcessingEnterprise Knowledge Search NVIDIA AI Enterprise · NVIDIA DGX A100 Tensor Core GPUs · NVIDIA NIM