The DPP, the grievance rail, and AgroFutures are three applications on one delivery layer. This page is the technical breakdown for the reader who wants to know how a compliance record, a farm advisory, and an abuse report all move over the same 2G signal.
Billions are being spent on AI equity and accessibility, and the unsolved bottleneck is the same everywhere: smartphone dependency, data dependency. Nobody has shipped a frontier model that answers on plain 2G. We did.
The Rail
A delivery layer that runs on 2G, on any carrier network, in any country: USSD and voice with local-language text-to-speech, plus a patent-pending platform for delivering AI knowledge over any SIM: no app, no data plan, no smartphone. It works where every other AI interface fails.
The reasoning layer runs any capable model on the backend. The rail is the reusable, hard-to-replicate part.
The Reasoning Layer
Every verified input becomes permanent infrastructure for everyone else on the rail.
A smartphone submitting a photo or a chat message isn't just getting an answer. The system checks that input against known conditions for the specific farm it came from, confirms or rejects the diagnosis, and writes the verified result into the same dataset every advisory on the rail draws from, including the ones delivered by voice callback to a farmer who has never held a smartphone.
The richer channels don't get a better product. They train the one system that serves everyone.
The Channel Spectrum
Any phone · 2G, no data
USSD menus, SMS, voice callback. Advisory, registration, reports, logging. Zero data cost.
Feature phone · with SIM
SMS plus voice IVR: outbound spoken advisory via local-language text-to-speech, interactive voice menus.
Smartphone · with data
WhatsApp and Telegram: free-text AI chat, photo diagnosis, farm-fused advisory, spoken voice notes. Same brain, richer screen.
Desktop / tablet · browser
Web dashboards for officers, auditors, and logistics operators reading the same data layer.
2G
Any carrier network
~1–2KB
Total session data per advisory
$15
Entry-level phone, the floor this rail is built for
The USSD and SMS architecture is live. WhatsApp and Telegram channels are live for AgroFutures. Our next-generation delivery layer, which removes the last carrier dependency, is in active development. Partnership and research inquiries welcome.
Five separate barriers stop a household from using a smartphone, and clearing four doesn't help if the fifth still stands. The rail is designed for the lowest common denominator device: it clears all five at once.
Rural Kenya sits at 25% internet usage against Nairobi's 64.7%. A national average is just two cities pulling the number up. The same five gates that stop a farmer in Murang'a stop a senior in Cleveland who cleared cost and supply years ago but never cleared literacy.
Sources: GSMA (via Forbes, Feb 2026) for the 3.4B / 3.1B / $3T figures; Pew Research / Eskaton via CITRIS Health for the 73% U.S. senior figure; CA Kenya / KNBS joint survey for the Nairobi/rural Kenya split.
A data desert isn't empty because nothing is happening in it. It's empty because nothing gets recorded. If you can SMS a chatbot, you can SMS a database. The same 2G connection that delivers a farm advisory can also capture registrations, pickup logs, labour hours, and abuse reports. Every upstream log feeds back into the intelligence layer, so the rail gets more useful the more it runs.
The channels aren't parallel tracks either. They feed each other. A WhatsApp photo from one co-op lead becomes an expert-reviewed advisory for a thousand 2G farmers who could never have taken that photo themselves. Every rich input from a smartphone (photo diagnosis, free-text chat, location share) enters the same knowledge layer that basic-phone farmers read from. Smartphone users are the sensor network. Everyone else receives what the sensors collected.
The delivery layer is available as a custom build for other organisations reaching populations across the full device spectrum, from a $15 phone on 2G to a smartphone on WhatsApp.
Work with us →