MNRE Launches AI-Based RAG System Assessment Initiative For Government Knowledge Management

The Ministry of New and Renewable Energy (MNRE), Government of India, has launched a major initiative to assess Artificial Intelligence-based Retrieval Augmented Generation (RAG) systems available in the Indian market. Through an official Expression of Interest (EoI) published on May 20, 2026, the ministry aims to study advanced AI-driven information retrieval solutions that can improve the way government data and documents are managed and accessed.

MNRE currently handles a vast and continuously growing repository of knowledge assets, including policy circulars, scheme guidelines, financial records, implementation reports, and various official documents. At present, most information retrieval activities depend on manual searches and keyword-based systems. According to the ministry, these traditional methods are becoming insufficient due to the increasing scale and complexity of operations.

To address this challenge, MNRE is exploring the use of Retrieval Augmented Generation (RAG) technology. This AI-based system combines semantic search capabilities with large language models to generate accurate and context-aware responses. One of the key advantages of RAG systems is their ability to provide proper source attribution while delivering information, which can improve efficiency, transparency, and decision-making within government departments.

The ministry has clearly stated that the EoI is only a market assessment exercise intended to gather technical insights and understand the capabilities available in the industry. It is not a tender, procurement process, or commercial offer. MNRE also clarified that no contracts, financial transactions, or business relationships will emerge from this exercise. Participation in the EoI will not provide any special preference or advantage in future procurement opportunities, which will be conducted separately under the General Financial Rules 2017.

The assessment process will take place in three stages. The first stage will involve written submissions and capability scoring. Interested participants must be registered Companies or Limited Liability Partnerships (LLPs) in India and should already possess a working AI or machine learning product ready for deployment. Submissions will be evaluated on five major criteria: prior project experience, document handling capacity, support for multiple document formats, multilingual capability, and deployment flexibility. The total score will be out of 100 marks.

Based on the evaluation, up to 15 top-scoring entities will be shortlisted for Stage 2, known as Independent Interface Testing. In this phase, selected vendors must provide access to a working system through a URL or sandbox environment within five working days of notification. MNRE’s internal testing committee will independently upload 250 to 500 public documents and conduct real-time testing without vendor involvement. The testing will assess system accuracy, multilingual performance, source citation quality, and handling of multiple document formats.

Stage 3 may include informal discussions with shortlisted participants to help MNRE better understand system architecture, deployment models, and operational factors. Interested companies must also disclose details related to data localization plans in India, client data protection during LLM training, and compliance with Government of India standards such as GIGW 3.0 and MeitY guidelines. Responses must be submitted in PDF format via email by 11:00 PM on May 27, 2026.