Back To Projects
CASE STUDY

Ministry of Social Development, Oman -Bilingual Policy Assistant

Arabic-English RAG chatbot with Arabic Sign Language (ArSL) recognition for inclusive citizen access to government policies. Deployed on-prem with full data residency.

MOSD Oman
MOSD Oman
Globcom
Globcom

Impact

The platform significantly improved how citizens access information, especially in cases where understanding policy language was previously difficult. Hearing-impaired users gained a new channel for interacting with government services. Citizens now receive faster responses, fewer queries require manual intervention, and the Ministry benefits from a more scalable support system.

1 / 2

Overview

The Ministry of Social Development needed an intelligent system that could help citizens quickly understand policies, access service information, and reduce dependency on manual support channels. The goal was a multilingual platform with strong accessibility support, including Arabic Sign Language (ArSL) input, and the ability to escalate to human assistance when required.

The Challenge

MOSD needed a single conversational system that could address multiple goals at once. Citizens needed easier access to policy information, guidance on social services, and help understanding government processes. The Ministry also wanted to reduce the burden on human support teams without lowering service quality. Key challenges included:

  • Large volumes of recurring queries
  • Complex policy documents across multiple categories
  • Limited accessibility for hearing-impaired citizens
  • Need for bilingual support and accurate retrieval
  • Requirement for secure, on-premises deployment
  • A future-ready architecture that can scale with new services

Objectives

  • Provide bilingual conversational access in Arabic and English
  • Reduce manual support load for common queries
  • Improve accessibility, including Arabic Sign Language input
  • Deliver high accuracy through a reliable RAG pipeline
  • Enable real-time tracking of service requests through MOSD APIs
  • Maintain data residency and security with on-prem hosting
  • Offer human escalation with both text and video call options

Our Approach

MythyaVerse engineered a complete end-to-end system combining multilingual RAG, sign language recognition, real-time service integration, and a user-centric conversational interface. The process began with mapping MOSD's policies, schemes, FAQs, guidelines, and public service definitions. Documents were cleaned, indexed, and prepared for retrieval with emphasis on accuracy and consistency. All components were deployed fully on-premises to meet data residency and security requirements.

Solution Delivered

  • Multilingual RAG Chatbot -High accuracy answers for policies, schemes, public services, and citizen queries in both Arabic and English.
  • Arabic Sign Language Input -Real-time ArSL recognition enabling hearing-impaired citizens to interact with government services.
  • Accessibility Focused UX -Clear layouts, simplified flows, and inclusive interaction paths designed for broad citizen use.
  • Human Escalation Layer -Built-in escalation with text and video call support for cases requiring personal assistance.
  • Integration with MOSD Services -Direct API connections allow citizens to track ongoing service requests from within the chatbot.
  • Unified Dashboard -A console for MOSD staff to track escalations, monitor conversations, and review insights.
  • Secure On-Prem Deployment -All components hosted within MOSD infrastructure with strict data residency guarantees.

Technologies Used

Next.jsNext.js
FastAPIFastAPI
DockerDocker
KubernetesKubernetes
PostgreSQLPostgreSQL

Explore More Projects

MythyaVerse - Building the Future