Sarvam Takes Indus to 2,500 Maharashtra Govt Officials

Sarvam AI is expanding its presence in India’s public sector after Maharashtra approved the deployment of its Indus AI workspace across state government departments. Around 2,500 government officials are expected to use Indus for tasks including policy research, drafting, document analysis, translation and data analysis, marking one of the larger deployments of an Indian-built AI system within a state administration.

The deployment is part of a broader two-year, ₹11.26-crore package approved by the Maharashtra government for Sarvam AI’s sovereign artificial intelligence solutions. The initiative also includes voice-based citizen outreach through Samvaad and speech technologies for the state’s MahaAI platform. The government says the systems will run on secure infrastructure located entirely within India, keeping official data inside the country.

Maharashtra Approves Sarvam AI Deployment

The Maharashtra government has approved a two-year outlay of ₹11.26 crore for Sarvam AI’s AI solutions across government departments.

The proposal was cleared by a High-Power Committee headed by the state’s chief secretary. Indus represents the largest component of the deployment and is expected to be made available to approximately 2,500 government users.

The first-year indicative cost for Indus is ₹5.63 crore, with the overall two-year allocation reaching ₹11.26 crore.

Deployment Details	Information
AI company	Sarvam AI
Main platform	Indus
Government users	Around 2,500
Total approved outlay	₹11.26 crore
Duration	Two years
First-year indicative cost	₹5.63 crore
Languages supported	22
Hosting	Infrastructure within India

The scale of the deployment is notable because government AI projects often begin as limited pilots within individual departments. Maharashtra is instead moving toward a broader deployment covering officials across multiple state departments.

What Indus Will Do for Government Officials

Indus is designed to assist officials with everyday knowledge and document-heavy tasks.

The platform can help with drafting and file noting, generate first drafts of responses, summarize documents and assist with policy research. Officials can also use it for data analysis and project-related workspaces.

Another important capability is document intelligence. Indus can process scanned records using optical character recognition (OCR), allowing information contained in physical or scanned documents to become searchable and easier to analyze.

The system supports 22 Indian languages, including Marathi, which could be particularly useful for Maharashtra’s administration.

Key Indus Use Cases

The deployment covers a broad range of administrative workflows:

• Policy and precedent research
• Drafting government notes and replies
• File-noting assistance
• RTI response generation
• Document summarization
• Data analysis
• OCR for scanned documents
• Multilingual interaction
• Project workspaces
• Document intelligence
• Integration with eOffice and NIC email where API access is available

This positions Indus as an administrative productivity tool rather than simply a chatbot for government employees.

Data Sovereignty Is a Major Focus

One of the most important aspects of Maharashtra’s deployment is where the AI infrastructure is hosted.

According to the government’s approval, Sarvam’s solutions will operate on secure and audited infrastructure located entirely within India. Government data is therefore intended to remain inside the country rather than being processed or stored overseas.

This is particularly relevant for government departments that handle sensitive documents, citizen information and internal administrative communications.

Data sovereignty has become an increasingly important consideration as governments worldwide adopt generative AI.

Instead of sending official information to externally hosted AI systems, governments can seek greater control over infrastructure, data storage and access.

Sarvam’s Indian-Language Advantage

Sarvam AI has positioned itself around AI systems designed specifically for India’s linguistic diversity.

Its models are being developed under India’s broader sovereign-AI push and are focused on Indian languages and use cases. Government documents describe Sarvam as one of the organizations developing sovereign foundational AI models under the IndiaAI Mission.

For state governments, multilingual capability can be particularly important.

Administrative work often involves a combination of English and regional languages. Maharashtra’s decision to deploy a system that supports Marathi alongside other Indian languages could allow officials to interact with AI tools without having to conduct every workflow exclusively in English.

This could also make AI more accessible to officials working outside highly English-dominant administrative environments.

The Broader ₹11.26-Crore AI Package

Indus is only one part of Maharashtra’s agreement with Sarvam.

The wider package includes Samvaad, a voice and WhatsApp-based citizen outreach system. The government has also approved the use of Saaras and Bulbul, Sarvam’s speech technologies, for integration with the state’s MahaAI platform.

This means Maharashtra is experimenting with AI across both internal administration and citizen-facing services.

AI for Citizen Outreach

Samvaad is intended to support voice-based interactions with citizens.

The approved programme includes a pilot involving tuberculosis awareness, in which an AI agent can make calls to patients in local languages and analyse the quality of those interactions.

The use case demonstrates how India’s AI adoption is moving beyond text-generation tools.

Speech recognition, multilingual voice systems and automated outreach could potentially make government programmes easier to deliver at scale, particularly when citizens prefer interacting in regional languages.

Maharashtra Is Building a Larger AI Strategy

The Sarvam deployment fits into a broader AI push by the Maharashtra government.

The state established a dedicated Electronics, Information Technology and Artificial Intelligence Department in April 2026. Its responsibilities include AI policy, digital governance, data governance, cybersecurity and digital infrastructure.

Maharashtra has also published an AI Policy 2026 and announced plans for an advanced State Data Centre in Nagpur.

The state’s approach suggests that AI is increasingly being treated as part of government infrastructure rather than as an experimental technology.

The partnership with Sarvam could therefore become a test case for how Indian states deploy domestically developed AI across large administrative organizations.

Sarvam Gains a Major Enterprise and Government Use Case

For Sarvam AI, the Maharashtra deployment provides more than a new customer.

A government-wide deployment involving thousands of officials can generate real-world feedback on how AI performs in document-heavy, multilingual and administrative environments.

Government workflows can also be significantly different from those of typical commercial businesses. Officials may need to work with lengthy policy documents, regulations, government orders, historical records and structured administrative procedures.

Successfully handling these requirements could strengthen Sarvam’s position in India’s enterprise and public-sector AI market.

India’s Sovereign AI Push Gains Momentum

The Maharashtra deployment also reflects India’s broader attempt to develop domestic AI capabilities.

The IndiaAI Mission has supported organizations working on foundational models, with Sarvam among the companies selected to develop sovereign AI technology. Government documents say the resulting models are intended to contribute to India’s AI ecosystem and are being made available through platforms such as AIKosh for developers and researchers.

The government’s objective is not simply to create an Indian alternative to foreign AI models.

It also aims to build domestic capabilities across models, computing infrastructure, datasets, applications and AI governance.

A large government deployment provides an important real-world application for that ecosystem.

Competition With Global AI Platforms

Sarvam’s expansion comes as Indian government organizations increasingly evaluate AI systems from both domestic and international companies.

Global AI companies have significant advantages in model scale, research resources and commercial reach. However, sovereign AI platforms can offer other benefits, particularly around data residency, local-language support and customization for Indian public-sector requirements.

Maharashtra’s decision to use Sarvam therefore highlights a potential market where local AI companies can compete on factors beyond raw model performance.

For government customers, the ability to keep data within India and support regional languages may be as important as benchmark scores.

Challenges Ahead

Despite the potential benefits, deploying generative AI across government departments comes with significant challenges.

AI-generated information can contain errors, meaning officials will still need to verify important outputs before using them in official decisions or communications.

Government documents can also contain sensitive information, making access controls, audit mechanisms and cybersecurity essential.

Another challenge will be measuring productivity gains.

The success of the programme will ultimately depend on whether Indus can reduce the time officials spend on repetitive administrative work without introducing new risks or additional verification burdens.

Training government employees to use AI responsibly will therefore be as important as providing access to the technology.

Looking Ahead

Maharashtra’s decision to deploy Sarvam AI’s Indus across government departments represents a significant step in the adoption of sovereign AI in India’s public sector. With around 2,500 officials gaining access to tools for research, drafting, document intelligence, OCR, translation and data analysis, the programme could provide one of the clearest large-scale tests yet of an Indian-built AI system in government workflows.

The larger significance extends beyond Maharashtra. If the deployment demonstrates measurable improvements in administrative productivity while maintaining data security and accuracy, it could encourage other state governments and public institutions to adopt similar systems. For Sarvam, the partnership could establish government as an important market for its multilingual and sovereign AI technology, while strengthening India’s broader effort to build AI infrastructure and applications under domestic control.