Digital Agriculture Mission / AgriStack — the corpus's strongest state-DPI agrifood AI anchor

South-Asia (India; pan-Indian; multi-state implementation)

Digital Agriculture Mission / AgriStack

Content

The Digital Agriculture Mission (DAM) was approved by the Government of India Cabinet on September 2, 2024, with a total outlay of ₹2,817 crore (~$321M) and ₹54.97 crore ($6.27M) allocated for FY 2025-26. The mission’s foundational component is AgriStack — a Digital Public Infrastructure (DPI) layer consisting of decentralized registries including the farmer registry, location-tagged plot registry, and crop-sown registry.

The unit’s structure is unusual: it is both deployment and framework. The deployment component (the four AI-enabled tools below) is operational at S3 scale; the framework component (AgriStack as a DPI architecture) is what the World Bank Connect4Impact paper (Feb 2026) treats as exportable as a model to other countries.

Deployed AI-enabled components (per PIB AI-and-Agriculture Explainer, Feb 14, 2026):

1. Kisan e-Mitra (AI chatbot)

   • Voice-enabled, AI-powered chatbot; answers queries on PM Kisan Samman Nidhi, Kisan Credit Card, Pradhan Mantri Fasal Bima Yojana
   • 11 regional languages
   • >93 lakh (9.3M) queries answered as of December 2025
   • >8,000 farmer queries/day
   • Launched 2023

2. National Pest Surveillance System

   • Supports 66 crops and 432+ pest types
   • Real-time advisories to >10,000 extension workers
   • Early pest detection

3. AI-based monsoon onset forecasting pilot (Kharif 2025)

   • Reached 3.88 crore (38.8M) farmers across 13 states via SMS
   • 31–52% of surveyed farmers adjusted sowing and land preparation decisions based on the forecasts

4. YES-TECH, CROPIC, PMFBY WhatsApp Chatbot

   • AI-enabled tools for crop insurance under PMFBY (Pradhan Mantri Fasal Bima Yojana)

AgriStack scale (per PIB Feb 14, 2026 explainer and Biometric Update coverage, Feb 2026):

• >7.63 crore (76.3M) Farmer IDs generated
• 23.5 crore (235M) crop plots surveyed
• Designed as pan-Indian DPI; rollout is in phases

Companion platform — Bharat-VISTAAR (announced Union Budget 2026-27, Feb 2026):

• “Virtually Integrated System to Access Agricultural Resources”
• Multilingual AI tool integrating AgriStack portals and ICAR package on agricultural practices
• Voice-first: dial 155261
• Initial rollout Hindi + English; multilingual expansion planned
• L0 announced at scan time (no deployment-scope figures surfaced)

Companion platform — Microsoft AI Sowing App / ICRISAT (referenced from a separate cycle unit if added):

• Pre-DAM research-to-deployment anchor, deployed since 2017 in Andhra Pradesh, Karnataka, Maharashtra, Telangana
• Voice-based via SMS / automated voice calls to feature phones

Companion platform — Farm Again (Tamil Nadu; cited in PIB Feb 14, 2026):

• AI-enabled precision farming with solar-powered sensors
• 3,500 farmers across >4,000 acres in Tamil Nadu
• Indigenous equipment cost ₹2.5 lakh per unit vs ₹25 lakh imported
• Annual savings: >4,00,000 cubic metres water, ~1,75,000 kWh energy, ~20,000 tonnes CO₂-equivalent emissions avoided
• Expansion to multiple countries claimed by the company

Institutional ecosystem:

• DA&FW (operational lead)
• MeitY (technology lead, IndiaAI Mission co-anchor)
• NITI Aayog (National Strategy for AI priority framing)
• ICAR and KVKs (research and extension)
• State agriculture departments
• Common Service Centers (CSC-SPV)
• Public-private partners (Microsoft, Cropin via ADPC, ITCMAARS)

What this unit is doing in the taxonomy

This unit is the state-DPI anchor for the corpus’s South-Asia region coverage and the corpus’s first deployment-scale state-stewarded data governance unit. It exercises:

• Sector positions 1 and 7 — on-farm-production-open-field (advisory, pest surveillance, monsoon forecast) and distribution-and-retail (AgriStack as registry layer for downstream services).
• AI technique classes 1, 5, 7 — predictive ML (yield forecasting, pest surveillance, monsoon onset), generative AI / LLMs (Kisan e-Mitra, Bharat-VISTAAR), sensors and IoT ML (Farm Again solar sensors, CROPIC).
• Purpose 3, 4, 9, 6 — climate adaptation (monsoon forecast, pest surveillance), supply chain efficiency (AgriStack registry as substrate for downstream services), food security / sovereignty (pan-Indian DPI frame), governance (data governance, traceability for state purposes).

It carries a maturity-scale S3 grade: 76.3M farmer IDs exceeds the S3 threshold. Maturity-verification V0: all figures are state-reported (PIB explainer); no independent audit cited. Maturity-longevity L1: DAM launched Sept 2024; 18 months operational at scan time; first-generation. Maturity-translation T3: institutional adoption pathway via DA&FW + MeitY + state agriculture + ICAR + KVKs + 10,000+ extension workers — a substantial institutional structure.

Why it matters for talks

Three reasons:

1. It is the corpus’s first deployment-scale state-DPI unit. The data-governance tag (state-stewarded) and data-rights-framework tag (state-stewarded) are now anchored in a real, named, scaled deployment — not a policy declaration. A talk on state-DPI vs vendor-owned vs cooperative data governance can use this unit as the canonical state-DPI case.
2. The World Bank Connect4Impact DPI paper (Feb 2026) treats India’s agrifood DPI as exportable as a model — a structural observation worth surfacing. The unit pairs naturally with indiaai-mission-agrifood.md (the policy frame) and with cropin-india.md (the private-platform layer) to name the three layers of India’s agrifood AI architecture.
3. The Kisan e-Mitra deployment figure (9.3M queries, 11 languages, 8,000 queries/day) is the corpus’s largest single-state-AI-deployment chatbot footprint. Worth surfacing in talks about voice-first / multilingual AI in low-literacy contexts — the deployment pattern is materially distinct from corporate chat assistants.

Critical context

• Vendor-figure hygiene: all scale figures are state-reported. PIB is tier-1 state source but the figures are not independently audited. V0 per corpus hygiene. Cross-reference G-070 for independent verification gap.
• State-stewarded data governance: this is a substantive digital-equity question. Per the TCI Cornell blog (Sept 2025) and UNESCO normative instruments, the farmer-registry data layer is state-stewarded but the downstream uses are not all transparent. The Internet Freedom Foundation’s “Agristack — a primer” raises consent / data sovereignty concerns. The corpus preserves the critical-voice tag (food-sovereignty) accordingly.
• Bharat-VISTAAR is announced but L0; the unit does not double-count the announced figure with deployed scale. The companion platforms (Microsoft ICRISAT, Farm Again) are referenced for context, not double-counted.
• ITCMAARS alignment: ITC has explicitly framed its platform as aligned with Bharat-VISTAAR / AgriStack direction. The two units (ITCMAARS and DAM) are paired but distinct — ITCMAARS is conglomerate-orchestrated, DAM is state-DPI.
• The cross-reference to the Microsoft / ICRISAT AI Sowing App: this is a research-to-deployment counterpart and predates DAM by ~7 years. If a cycle adds that unit, it should be cross-referenced.