A definitive product, technology, and growth strategy — written to win investor confidence, survive real rural execution, and outperform every leading agritech competitor without losing simplicity.
The thesis. India has 146 million smallholder farmers losing ₹2.4 lakh crore every year to bad weather guesses, mistimed selling, failed insurance claims, and predatory credit. India's public digital stack — AgriStack, ONDC, OCEN, Bhashini, DigiLocker, AGMARKNET — quietly matured between 2024 and 2025 and made it possible, for the first time, to build a farmer-side operating system without spending billions on infrastructure. AgriSense AI is that operating system. Voice-first in 15+ Indian languages. Free to the farmer forever. Monetised by insurers, lenders, marketplaces, and state governments. Already live with 8,742 pilot farmers.
Why this beats every competitor. Nobody has all four pillars: open-stack integration, voice-first multilingual UX, free at the base, B2B revenue away from the farmer. On a 14-capability scorecard, AgriSense scores 14/14 vs 6/14 for the next best.
The 2030 ambition. 100K farmers in Year 2 → 5M in Year 4 → 50M in Year 7 — the agriculture intelligence layer for the entire Global South.
The ask. ₹4.5 crore pre-seed for 18 months. 35% engineering, 30% FPO ops, 20% govt BD, 15% reserve.
Core philosophy. Every new feature must pass three gates: (1) Does it speak Telugu? (2) Does someone other than the farmer pay for it? (3) Does it work offline on a ₹6,000 Android?
Replace today's daily card with a two-way Telugu voice agent (Llama-3-8B distilled + LoRA on 5k Telangana voice samples).
One Aadhaar = one head-of-household account; spouse + parents + daughter linked.
Daily action card pushed as Telugu voice note to existing WhatsApp users — meet farmers where they already are.
Single home-screen button: "Crop damaged today?" → 3 photos → drone dispatch → claim filed in 42 minutes.
Weekly scan against 200+ scheme rules-engine; pings farmer when new scheme matches.
Aggregate 8-10 farmers in 2 km → drone spray cost ₹780/ac → ₹220/ac.
Three lender quotes in 60 seconds based on AgriStack + 12 months platform activity.
Verra-certified regenerative practices · farmer keeps 75% of $40-80/ton value.
Scored across 14 capabilities. Full = 1, partial = 0.5.
| Player | Score / 14 | Closest threat? |
|---|---|---|
| AgriSense AI (us) | 14.0 | — |
| DeHaat | 6.0 | Only real India threat over 18-24 month window — could pivot |
| Cropin Technology | 5.5 | Enterprise SaaS only — doesn't serve farmer |
| Apollo Agriculture (Kenya) | 5.5 | Different geography · model match |
| AgroStar | 4.0 | Input commerce only · missing AI advisory |
| Plantix (PEAT) | 3.5 | Disease classifier only · misses 16 of 17 problems |
| Climate FieldView (Bayer) | 3.5 | US/EU built · smallholder UX unsolved |
| Ninjacart | 3.0 | B2B supply chain · farmer is supplier not user |
| Fasal | 2.0 | Horticulture IoT premium only |
| BigHaat | 2.0 | Input e-commerce only |
| BharatAgri † | 1.5 | CEASED OPERATIONS Nov 2025 |
RICE = (Reach × Impact × Confidence) ÷ Effort. Top-4 ship order: PMFBY filer → Auto-subsidy radar → WhatsApp agent → Drone convoy.
| Feature | Reach | Impact | Conf. | Effort | RICE | When |
|---|---|---|---|---|---|---|
| One-tap PMFBY filer | 100k | 5 | 95% | 2 | 237,500 | Q3 2026 |
| Auto-subsidy radar | 100k | 3 | 95% | 2 | 142,500 | Q3 2026 |
| WhatsApp Business agent | 80k | 4 | 90% | 3 | 96,000 | Q3 2026 |
| Family accounts | 60k | 3 | 90% | 2 | 81,000 | Q4 2026 |
| Conversational Telugu copilot | 100k | 5 | 80% | 6 | 66,667 | Q4 2026 |
| OCEN credit on tap | 50k | 5 | 85% | 4 | 53,125 | Q4 2026 |
| Drone convoy-mode | 40k | 4 | 90% | 3 | 48,000 | Q3 2026 |
| Carbon credit issuance | 25k | 4 | 70% | 5 | 14,000 | Q1 2027 |
100K farmers · TS+AP+KA · 2 State DaaS · 500 FPOs · ₹15 cr ARR · Series A close.
5M farmers · 18 states · smart-village layer · carbon credits live · APIs open to 3rd parties · ₹150 cr ARR.
50M farmers · Pan-India + Bangladesh + Sri Lanka + East Africa · ₹740 cr ARR · IPO or strategic exit.
Five streams. The farmer never pays for anything essential.
| Stream | Customer | Pricing | Y1 | Y3 | Y7 |
|---|---|---|---|---|---|
| Insurer Evidence API | AIC, ICICI Lombard | ₹120/claim | 0.8 cr | 12 cr | 85 cr |
| Lender Lead-Gen | HDFC, ICICI, NABARD | ₹450/₹1L | 0.4 cr | 8 cr | 60 cr |
| District-as-a-Service | State govt | ₹35-80 L/dist | 2.5 cr | 38 cr | 280 cr |
| Marketplace take-rate | Input/output (ONDC) | 1.2-2.5% | 0.6 cr | 14 cr | 110 cr |
| Premium subscription | ~12% farmer base | ₹99-299/mo | 0.3 cr | 6 cr | 70 cr |
| Carbon + ESG (Y3+) | Verra buyers | $40-80/ton | — | 4 cr | 95 cr |
| API monetisation (Y4+) | Developers | Per-call | — | 2 cr | 40 cr |
| TOTAL ARR | ₹4.6 cr | ₹84 cr | ₹740 cr | ||
React Native (11 MB · Android 8+) · quantized on-device ML · IVR/USSD/SMS via Twilio/Exotel/Mitra · IoT (ESP32 moisture, DHT22 warehouse, telematics).
AgriStack · ONDC (Beckn) · OCEN · DigiLocker · AGMARKNET · Bhuvan · Bhashini · AIC PMFBY · NPCI UPI Mandate Pay.
Node + Python microservices on K8s (AWS Mumbai + Azure Hyderabad) · Postgres + ClickHouse · Kafka events · KServe ML serving · Cloudflare R2.
| Model | Role | Runs on |
|---|---|---|
| LightGBM + GraphCast | 1km × 30min rain nowcast | Server |
| EfficientNet-V2 (INT8) | 65 crop diseases from photo | Edge + server |
| LSTM | Pump health · cold-chain alerts | Server |
| Bhashini ASR + TTS + NMT | 15+ Indian languages | Server (GPU) |
| Llama-3-8B + LoRA | Telugu conversational agronomy | Server |
| Custom credit-score ensemble | OCEN underwriting | Server |
| YOLOv8 | Drone NDVI/thermal damage | Edge (drone) + server |
| ARIMA + Prophet | 30-day mandi price band | Server |
| Vision Transformer | Sentinel-2 biomass tracking | Server |
| Verra MRV ensemble | Carbon sequestration measurement | Server |
Three design laws, never broken:
Text is the exception. Every screen has a "tap to listen" button. Home shows one daily card with one action verb.
From app launch to "I know what to do today" is no more than 3 taps. Most farmers reach value in 2.
Network drops → cached card still shows. Voice fails → 3-button menu. Phone breaks → SMS keeps working.
Onboarding flow (60-90 seconds): Sarpanch enters Aadhaar + mobile → auto-pull AgriStack → app or IVR auto-registers → welcome voice call within 24h → first daily card Day 2 → first small win Week 1.
Every integration is a separately monetisable surface AND a defensibility moat.
Identity + land → onboarding ₹14 (vs ₹1,200 industry)
Buy + sell rails → marketplace take-rate
Loan rails → ₹450/lakh lender fee
Scheme application auto-fill
Live mandi prices
15+ language ASR/TTS/NMT
₹120/claim insurer revenue
₹35-80 L/district/yr DaaS contracts
Working MVP · 8,742 pilot users · public-stack thesis unique · ₹38L pilot burn · ESG-aligned · BharatAgri shut down clears competitor.
ESG capital pool growing · ONDC+AgriStack maturing 18-24 month FMA window · funding redirected from BharatAgri.
DeHaat pivot (sign 6-8 state DaaS first) · policy risk (rail-agnostic abstraction) · copycat (data moat & community moat).
| Moat | Time to replicate |
|---|---|
| Data moat — 2M+ Telugu voice + disease photo + farmer interaction records | 18-24 mo minimum |
| AI moat — Bhashini fine-tunes, Telugu vocabulary, regional disease weights | 12-18 mo |
| Distribution moat — 47 FPO contracts + 2 state DaaS + bank channels | 24+ mo (relationship-led) |
| Community moat — farmer evangelists in 500+ villages (k=1.27) | Impossible without ground presence |
| Integration moat — 12 public-stack integrations in production | 18-30 mo per competitor |
Asset-light by design. AgriSense owns no warehouses, drones, tractors, or inventory. We orchestrate on top of partners. Pilot ran on ₹38 lakh; Twiga (capex-heavy) collapsed in 2025.
| Metric | 10K farmers | 100K | 5M |
|---|---|---|---|
| CAC (blended) | ₹14 | ₹16 | ₹22 |
| LTV (Year 1) | ₹680 | ₹720 | ₹980 |
| LTV (Year 3 cumulative) | ₹6,800 | ₹8,200 | ₹14,200 |
| Infra cost / farmer / month | ₹4.20 | ₹2.80 | ₹1.40 |
| Gross margin | 72% | 78% | 84% |
| EBITDA margin | (28%) | 12% | 38% |
| State | Wedge | Crop focus | Y2 target |
|---|---|---|---|
| Telangana | Existing pilot · Sarpanch + FPO depth | Paddy, cotton, chilli | 30,000 |
| Andhra Pradesh | RYTHU BHAROSA overlay | Paddy, mango, tomato | 25,000 |
| Karnataka | Horticulture-first (Fasal-style) | Grape, pomegranate, areca | 15,000 |
| Maharashtra | Sugarcane mills + coop bank | Sugarcane, soybean | 18,000 |
| Punjab | Wheat/paddy + PSPCL rebate | Wheat, basmati | 8,000 |
| Tamil Nadu | Coconut + millet diversification | Coconut, millet | 12,000 |
50M Indian farmers + families on one voice-first card.
Global South layer · 200M farmers · BD, LK, NP, KE, TZ, ID.
Drone fleets + sensors + smart pumps coordinated by AgriSense.
$1.5-3B issued annually · farmer keeps 75%.
De-facto smallholder credit bureau of India.
ONDC's most active smallholder-side aggregator · ₹40,000 cr GMV.
8 years of micro-climate + outcome data for govt early-warning.
50M farmers · ₹740 cr ARR · IPO-ready.
"Free for the farmer, paid for by everyone who profits when the farmer succeeds."
This is not a demo project. This is the AI Agriculture Operating System for the next billion farmers — and we are going to build it whether you fund us or not. The only question is whether you want to be on the cap table when it becomes the rails of Indian agriculture.
📄 Download full Strategy 2030 PDF View investor pitch