AI for the people
the smartphone
economy left behind.

PHLA · LONDON · NEW YORK · NAIROBI
↓
The Problem

ALL KNOWLEDGE IS ACCESSIBLE.

The infrastructure is built to give the same AI systems to users on 2g phones to the CEO in NYC. We developed the intelligence layer to bring this knowledge to this that are being left behind at an ever increasing rate.

3.4B
People without mobile internet
GSMA via Forbes, Feb 2026 ↗
3.1B
Live under signal. Just can't use it.
GSMA "Usage Gap" via Forbes ↗
$3T
Annual cost to global GDP
GSMA via Forbes, Feb 2026 ↗
73%
Of low-income U.S. seniors without a smartphone — the same figure as rural Kenya
Pew Research / Eskaton via CITRIS Health ↗
$406B
Annual U.S. cost of senior isolation
NYSOFA / CDC estimate ↗
95%
Loneliness reduction — AI companion pilot
NYSOFA + ElliQ Report, 2023 ↗ · 2026 update ↗

The share of low-income U.S. seniors without a smartphone is statistically identical to the share of rural Kenyans without one. The digital divide is not a developing world problem. It is a poverty problem. We built one solution for both.

East Africa — Agriculture
A fisherman in Lamu needs to know if it's safe to go out.

96% of rural East African farmers use basic feature phones. Not by choice — because a $15 Nokia outlasts a smartphone in dust and heat, costs nothing to insure, and works on 2G where 4G doesn't reach.

Every AI agricultural advisory platform ever built requires a smartphone, a data plan, and digital literacy. AgroFutures requires none of them. A farmer dials a shortcode, selects their crop, and receives a real-time advisory — generated from satellite data, biological pest models, and live weather — read aloud in their language via automated voice callback.

The intelligence is frontier. The phone is $15.

300M addressable users · East Africa alone
United States — Senior Health
A widow in the Bronx hasn't spoken to anyone in three days.

Loneliness kills. The U.S. Surgeon General's 2023 advisory found that lacking social connection increases the risk of premature death by more than 60% — comparable to smoking 15 cigarettes a day.

New York State deployed 900 AI companion robots to address this. The results were extraordinary: 95% reduction in loneliness, 30+ daily interactions per user. But robots are expensive. County rollouts are slow. And none of it reaches the 30 million seniors who don't own a smartphone.

They all have phones. They all know how to text.

$6.7B annual Medicare cost from isolation alone
The Language Problem

AI speaks English fluently.
It guesses at everything else.

Seven thousand languages are spoken on earth. The major AI platforms were trained on a fraction of them. For the populations we serve — smallholder farmers in Kenya, coastal fishing communities, rural health workers — the language gap is not a nuisance. It is a complete barrier to access.

We moved to Khaya AI (GhanaNLP) for African-language text-to-speech, and are integrating PazaBench ASR models from Microsoft Research Africa for voice input in Dholuo, Kalenjin, Kikuyu, Maasai, and Somali. Current deployment: English, Kiswahili, Dholuo, Kikuyu. Expanding to 12+ languages via LINGUA Africa.

Language is not a localisation problem. It is an access problem.

Google Translate — Kikuyu input Identified Kikuyu as a Congolese Creole. Translation was structurally incoherent.
Gooey AI — Kikuyu agricultural query Translated a regional coffee variety name as a personal name. Advice was nonsensical.
Why it matters A farmer who receives a mistranslated pest advisory doesn't just miss the information. They may act on the wrong one.
How We Got Here
01
USSD

We started with USSD — the protocol that runs *384# menus on any phone on any network. We built a bilingual advisory system for farmers in East Africa: no smartphone, no data, no app. It worked. We proved the thesis, hit the ceiling, and open-sourced the architecture.

02
SIMkit

The real breakthrough wasn't the advisory. It was realising the delivery layer was still borrowed. USSD depends on carrier gateways that fail, rate-limit, and charge per session.

So we went deeper. The SIM Toolkit (STK) is a native Java Card applet that lives inside the SIM itself — the menu that ships inside every phone on earth, on every network, in every country. Silent SMS as the transport layer. Cloudflare Edge Workers for routing. The phone never talks to the AI directly. It talks to a shortcode.

No gateway. No app. No middleware. No data plan required.

03
What SIMkit Became

Once the transport layer was solved, the question changed. It was no longer "how do we reach farmers?" It became: who else does this reach?

The same architecture that delivers a crop advisory in Dholuo to a farmer in Murang'a delivers a mental health check-in to an isolated senior in the Bronx. The same protocol handles women's health information in markets where smartphones are unaffordable and literacy is low. Healthcare. Civic services. Financial guidance.

SIMkit is not an app. It is infrastructure. The verticals are what we build on top of it.

Research Tracks

Two deployed applications. One platform underneath.

East Africa — Agriculture & Fisheries

AgroFutures

AI crop and marine advisories delivered over USSD and voice to feature phone users in East Africa. No smartphone. No data plan.

Deployed in Murang'a County (150,000 registered farmers), with coastal fisheries coverage in Lamu. Integrates satellite data (NDVI, SST, chlorophyll), biological pest models, and multilingual voice delivery via Khaya AI in Dholuo, Kikuyu, and Kiswahili.

DEPLOYED · LIVE

Microsite ↗ GitHub ↗
United States — Health & Social Care

Senior Health SMS

AI companionship and health support delivered over SMS to older adults on basic phones, addressing the senior isolation crisis.

30 million U.S. adults over 65 don't own a smartphone. They all have phones. They all know how to text. This gateway connects any SMS-capable device to a warm, context-aware AI companion — built for Area Agencies on Aging, home health aides, and senior centers running on shared hosting, not enterprise budgets.

OPEN SOURCE · IN DEVELOPMENT

GitHub ↗
The Platform

SIMkit

The SIM is the terminal.

SIMkit is a Java Card STK applet that turns the native SIM menu into an AI interface. Silent SMS as transport. Cloudflare Edge Workers for routing. Any LLM on the backend.

It runs on 2G. It requires no data plan, no app installation, no smartphone. It works in network-constrained environments where every other AI interface fails.

Current verticals: agriculture, senior companionship.
In development: mental health, women's health, general healthcare, civic services.

The architecture is the same across all of them.

2G

Any carrier network

~1–2KB

Total session data per advisory

3B+

Addressable feature phone users

SIMkit is in active development. Partnership and research inquiries welcome.

S

SAUTI NETWORKS — SIM Delivery Platform

Step 0 / 7
NOKIA
Select Back
OK
sauti-stk — protocol trace
swipe left to advance
Open Source

The delivery infrastructure is MIT licensed. We believe the protocols that connect excluded populations to AI should be owned by no one. The intelligence models, biological simulations, and advisory engines built on top of them are proprietary.

Fork these repos to build your own applications on 2G infrastructure. Build for your market, your language, your population.

africas-talking-agtech

USSD + voice advisory interface for East Africa. PHP. MIT.

GitHub ↗

twilio-senior-health-2g-AI-access

SMS AI companion gateway for older adults. PHP. MIT.

GitHub ↗

simkit

STK applet + silent SMS transport layer. In development.

Coming Soon
Contact

Work with us.

We work with researchers, NGOs, development agencies, and funders building toward AI access for excluded populations.

For partnerships, grant co-applications, or technical collaboration:

[email protected]

grafikinc.com