Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

KrishokBondhu: A Retrieval-Augmented Voice-Based Agricultural Advisory Call Center for Bengali Farmers

Mohd Ruhul Ameen, Akif Islam, Farjana Aktar, M. Saifuzzaman Rafat

TL;DR

KrishokBondhu presents a Bengali, voice-based advisory system built on Retrieval-Augmented Generation (RAG) to deliver expert agricultural guidance to Bangladeshi farmers. It grounding responses in a curated Bengali knowledge base through OCR, semantic retrieval with LanceDB, and generation by Gemma 3-4B, delivered via VAPI for speech interfaces. In a pilot, the system achieved 72.7% high-quality responses and an overall score of 4.53, outperforming the Kishan QRS baseline by 44.7%, with substantial gains in contextual richness and completeness. The work demonstrates the feasibility of integrating call-center accessibility, native-language voice interaction, and RAG techniques to democratize agricultural knowledge, while outlining future improvements like broader content coverage and real-time data integration.

Abstract

In Bangladesh, many farmers continue to face challenges in accessing timely, expert-level agricultural guidance. This paper presents KrishokBondhu, a voice-enabled, call-centre-integrated advisory platform built on a Retrieval-Augmented Generation (RAG) framework, designed specifically for Bengali-speaking farmers. The system aggregates authoritative agricultural handbooks, extension manuals, and NGO publications; applies Optical Character Recognition (OCR) and document-parsing pipelines to digitize and structure the content; and indexes this corpus in a vector database for efficient semantic retrieval. Through a simple phone-based interface, farmers can call the system to receive real-time, context-aware advice: speech-to-text converts the Bengali query, the RAG module retrieves relevant content, a large language model (Gemma 3-4B) generates a context-grounded response, and text-to-speech delivers the answer in natural spoken Bengali. In a pilot evaluation, KrishokBondhu produced high-quality responses for 72.7% of diverse agricultural queries covering crop management, disease control, and cultivation practices. Compared to the KisanQRS benchmark, the system achieved a composite score of 4.53 (vs. 3.13) on a 5-point scale, a 44.7% improvement, with especially large gains in contextual richness (+367%) and completeness (+100.4%), while maintaining comparable relevance and technical specificity. Semantic similarity analysis further revealed a strong correlation between retrieved context and answer quality, emphasizing the importance of grounding generative responses in curated documentation. KrishokBondhu demonstrates the feasibility of integrating call-centre accessibility, multilingual voice interaction, and modern RAG techniques to deliver expert-level agricultural guidance to remote Bangladeshi farmers, paving the way toward a fully AI-driven agricultural advisory ecosystem.

KrishokBondhu: A Retrieval-Augmented Voice-Based Agricultural Advisory Call Center for Bengali Farmers

TL;DR

KrishokBondhu presents a Bengali, voice-based advisory system built on Retrieval-Augmented Generation (RAG) to deliver expert agricultural guidance to Bangladeshi farmers. It grounding responses in a curated Bengali knowledge base through OCR, semantic retrieval with LanceDB, and generation by Gemma 3-4B, delivered via VAPI for speech interfaces. In a pilot, the system achieved 72.7% high-quality responses and an overall score of 4.53, outperforming the Kishan QRS baseline by 44.7%, with substantial gains in contextual richness and completeness. The work demonstrates the feasibility of integrating call-center accessibility, native-language voice interaction, and RAG techniques to democratize agricultural knowledge, while outlining future improvements like broader content coverage and real-time data integration.

Abstract

In Bangladesh, many farmers continue to face challenges in accessing timely, expert-level agricultural guidance. This paper presents KrishokBondhu, a voice-enabled, call-centre-integrated advisory platform built on a Retrieval-Augmented Generation (RAG) framework, designed specifically for Bengali-speaking farmers. The system aggregates authoritative agricultural handbooks, extension manuals, and NGO publications; applies Optical Character Recognition (OCR) and document-parsing pipelines to digitize and structure the content; and indexes this corpus in a vector database for efficient semantic retrieval. Through a simple phone-based interface, farmers can call the system to receive real-time, context-aware advice: speech-to-text converts the Bengali query, the RAG module retrieves relevant content, a large language model (Gemma 3-4B) generates a context-grounded response, and text-to-speech delivers the answer in natural spoken Bengali. In a pilot evaluation, KrishokBondhu produced high-quality responses for 72.7% of diverse agricultural queries covering crop management, disease control, and cultivation practices. Compared to the KisanQRS benchmark, the system achieved a composite score of 4.53 (vs. 3.13) on a 5-point scale, a 44.7% improvement, with especially large gains in contextual richness (+367%) and completeness (+100.4%), while maintaining comparable relevance and technical specificity. Semantic similarity analysis further revealed a strong correlation between retrieved context and answer quality, emphasizing the importance of grounding generative responses in curated documentation. KrishokBondhu demonstrates the feasibility of integrating call-centre accessibility, multilingual voice interaction, and modern RAG techniques to deliver expert-level agricultural guidance to remote Bangladeshi farmers, paving the way toward a fully AI-driven agricultural advisory ecosystem.
Paper Structure (14 sections, 7 figures, 5 tables)

This paper contains 14 sections, 7 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. System Architecture and Methodology
  4. Data Collection and Source Curation
  5. Document Processing and OCR Pipeline
  6. Vector Database and Retrieval System
  7. Language Model Integration and Response Generation
  8. Voice Interface Integration with VAPI
  9. Evaluation Methodology
  10. Results and Discussion
  11. Overall System Performance
  12. Comparative Analysis with Kishan QRS
  13. Semantic Similarity and Response Quality Correlation
  14. Conclusion

Figures (7)

  • Figure 1: Document Processing and OCR Pipeline
  • Figure 2: Vector Database and Retrieval Workflow
  • Figure 3: Voice Interface Integration Workflow with VAPI
  • Figure 4: Distribution of response quality categories across test questions, showing strong performance with 72.7% high-quality responses.
  • Figure 5: Radar chart comparing Krishokbondhu and Kishan QRS across five evaluation criteria, showing Krishokbondhu's superior performance in contextual richness and completeness.
  • ...and 2 more figures