Proof of Work

Real Projects. Proven Results.

We don't just talk — we deliver. Here are five representative projects we deployed for government and legal bodies in Israel.

Client details and sensitive data are kept strictly confidential — further details available under NDA

5+

government bodies & municipalities

20+

AI systems deployed

100%

client satisfaction

Level 5-6

security clearance

Government Ministry2025
24/7availability
< 3sresponse time
99%uptime
RAGGCPFlowiseHebrew NLP

Legal Information Chatbot for Public Access

A central government body needed to make a massive legal information database accessible to the public, while maintaining full factual accuracy. Questions arrive in Hebrew and Arabic, often in informal language.

Solution

We built a GCP-based RAG system with a hybrid query engine (semantic + exact), fine-tuned for Hebrew. The system answers only from official source documents and cites specific passages in every response.

Outcomes

  • Handles thousands of public queries daily without human agents
  • Over 95% factual accuracy — validated by manual review on a sample set
  • 24/7 availability with average response time under three seconds
Discuss a similar project
Judicial System2024–2025
2AI agents
10K+cases analyzed
Level 5security
AI AgentsAWS EC2PythonDocument AI

AI Case-Weight Assessment System for the Court

The Supreme Court needed an objective mechanism to determine and prioritize hearing assignments, while meeting strict government security standards. The challenge: assigning case weight requires reading, understanding and analyzing dozens of pages per case.

Solution

We built two specialized AI agents operating in sequence: Agent 1 extracts and structures relevant information from the case file; Agent 2 analyzes and scores according to an official weighting guide. All deployed on AWS EC2 within the organization.

Outcomes

  • Automated case analysis with structured weight report generated in minutes
  • Increased consistency and objectivity — zero subjective factor influence
  • Full integration with the existing case management system (Net HaMishpat)
Discuss a similar project
Government Ministry2025–2026
80%time reduction
Weeklycadence
Hebrewlanguage
Fine-TuningHebrew NLPLangChainOpenAI

AI-Assisted Linguistic Editing of Legislation

The editorial team responsible for Israeli legislative language faced the challenge of maintaining consistency, precision and strict linguistic rules across a high volume of legislation texts.

Solution

We implemented an AI tool combining fine-tuning on an Israeli legislative corpus with a rule-based checking mechanism. The tool identifies deviations, suggests corrections and explains each change according to the relevant rule.

Outcomes

  • Significant reduction in linguistic editing time per legislative body
  • Improved consistency — automatic detection of style deviations
  • Weekly status meetings with continuous feedback-driven improvements
Discuss a similar project
Municipality2025
500+documents/month
90%time reduction
4languages
Agentic AIn8nRAGDocument Processing

Automated Municipal Committee Protocol Processing

A large municipality generates hundreds of committee meeting protocols monthly. Manual transcription, summarization, decision extraction and filing consumed hours of work per meeting.

Solution

We built an end-to-end automated pipeline: audio file input → transcription → summarization → decision and task extraction → automatic filing. The system supports Hebrew, Arabic, Russian and Romanian.

Outcomes

  • 90% reduction in protocol processing time — from 3 hours to under 20 minutes
  • Automatic extraction of decisions and tasks with assignment to responsible parties
  • Full support for 4 languages — serving a diverse population
Discuss a similar project
National Agency2025
100+participants
2sessions
Nationalscope
AI AgentsOpenWebUIClauden8n

Agentic Workspace Workshop for Government Employees

A national government agency needed to train hundreds of employees in practical use of AI agents — not basic tools, but complex automated workflows. The team was diverse: managers, analysts, and developers.

Solution

We designed and delivered two intensive training sessions. Each participant left with a personal agentic workspace — a model router, integrated tools and a workflow tailored to their role. Methodology: learning by doing.

Outcomes

  • 100+ government employees trained to work with AI agents in practice
  • Every participant produced a working automation by the end of the workshop
  • The workshop served as the foundation for the national government AI agents committee
Discuss a similar project

Interested in more details?

Some of our projects include sensitive information not published publicly. We are happy to share additional details, architectures and full results in a direct conversation under NDA.

Book a Call

Explore More

Our Solutions

AI services for government and legal sectors

About Lawcal AI

Our team, vision, and clients

Contact Us

Book an AI consultation