From the ground up.

The delivery layer for reaching people no app can reach, on any phone, any network, any country.

From The Ground Up reaches them anyway: any phone, any network, any country, over 2G USSD, the same signal layer that already moves $2 trillion in mobile money a year. Reaching them does something no one else does. It creates a dataset that's never existed before, real signal, not a survey, not a guess, read straight off the channel people already use. We carry AI advisory, EU compliance evidence, silent abuse reports, and the first search-trend dataset an unconnected population has ever produced. No app. No data plan. No smartphone, for them or for whoever needs to finally see them.

The Problem

Access.

3.4 billion people don't use mobile internet. 3.1 billion of them live under a signal they can't use, the usage gap is nine times the coverage gap. Five separate barriers stop the rest: cost, data, literacy, language, power. Clearing four doesn't help if the fifth still stands.

Gate 01 · Cost
$3.5T
Potential GDP gain by 2030 from closing the usage gap alone (GSMA, Sept 2025). A smartphone that survives Mombasa's salt air fails faster than it pays for itself. Total cost of ownership, not sticker price.
Gate 02 · Data
3.1B
Live under signal, can't use it: GSMA's "usage gap" (GSMA, Sept 2025). A data plan in Kilifi can cost more per month than the maize it's meant to help sell.
Gate 03 · Literacy
73%
Of low-income U.S. seniors have no smartphone at all (Pew Research / Eskaton via CITRIS Health). Reading a screen and using one are two different skills.
Gate 04 · Language
800
Articles in Dholuo Wikipedia, against English's 6.8 million. Content exists; it doesn't exist in the language the reader speaks.
Gate 05 · Supply
300M
Have no mobile internet coverage at all, no towers, no signal, regardless of device (GSMA, Sept 2025). Parts of rural Uganda still aren't on the national grid. A device with no power to charge it doesn't reach anyone.

None of that is one problem with one fix. It's five separate gates, and a household can clear four of them and still be stopped by the fifth.

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. One rail, designed for the lowest common denominator device, clears all five at once.

Sources: GSMA (Sept 2025) for the 4.7B / 3.4B / 3.1B / 300M / $3.5T 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.

EU Compliance · Live

The farm can no longer author its own compliance evidence.

From December 30, 2026 (large and medium operators; June 30, 2027 for micro and small enterprises), regulated commodities entering the EU require verifiable due diligence data proving they weren't grown on deforested land. From 2029, in-scope EU importers must prove the workers who produced those goods weren't exploited. By ~2030, textiles.

Legacy compliance software collected that proof from the farm manager. The farm manager said yes. That's no longer sufficient. We collect it from three independent sources, none of them the farm:

Input 01 · The Worker
Logs their own hours via USSD, against their National ID. Any phone, any network. No employer device, no employer permission.
Input 02 · The Driver
Independently logs volume collected at the farm gate via USSD, cross-verified against the worker's record.
Input 03 · The Mobile Money Operator
Confirms the actual payout on its own ledger. Mathematical proof that wages were paid, in what amount, and whether they meet statutory and living-wage benchmarks.

Three independent 2G inputs. The farm authors none of them. No single party can falsify the record alone. The output is a Digital Product Passport covering CSDDD labour due-diligence, EUDR deforestation screening, and ESPR product traceability, one ledger, every rulebook.

See the live compliance portal → dpp.agro-futures.com
CSDDD Article 14 · Live

A complaint line no supervisor can switch off.

70–80% of workers on East African flower and tea farms are women. Investigative reporting by the BBC and Journalismfund Europe has documented the conditions: chemical exposure without protective gear, unpaid overtime, sexual harassment, and, in the words of one worker, no way to speak out.

The funded response is hotlines. Most run on WhatsApp. WhatsApp requires a smartphone the worker may not have, a data plan she may not afford, and her own device, which the employer can identify.

Our grievance rail runs on USSD. A worker borrows any phone in the village. Dials a shortcode. Files a complaint against her National ID. The reporter's identity never appears on any dashboard, only a case number and its status. The employer can't identify who filed. The employer can't block the channel. The employer can't confiscate the device, because it isn't hers.

An unresolved zero-tolerance grievance, physical abuse, child labour, flags the consignment as non-compliant on the buyer's dashboard, preventing a clean due diligence statement from being generated for any shipment linked to that farm.

The existing hotlines are reporting tools. This is a market-access mechanism.

Sources: BBC News, "Toiling on a Kenyan flower farm to send fresh roses to Europe" (Sept 2024); Journalismfund Europe, "An Unrosy Affair" (Jan 2024); GSMA Mobile Gender Gap Report 2026.

East Africa · Agriculture · Live

The intelligence is frontier. The phone is $15.

A farmer in Murang'a photographs a sick leaf on a co-op lead's phone via Telegram. The system doesn't just identify Fall Armyworm, it checks the diagnosis against his specific plot's humidity, growth stage, soil, and known pest pressure. When the conditions don't match the classifier's read, it hedges its own answer and asks for more evidence rather than forcing a false diagnosis. Then it speaks the advisory back in Dholuo, because he can't read it.

AgroFutures is a terrain-aware agricultural intelligence platform active across four climate zones: Mediterranean, tropical highland, tropical coastal, and aquaculture. It delivers per-plot AI advisory from a zero-data USSD line to WhatsApp and Telegram, with photo diagnosis, spoken voice notes in local languages, pest cluster early warning, and a Regional Portfolio Manager for the officer covering 1,000+ farmers.

One cent per advisory. Runs on any phone that can make a call.

USSD advisory flow, feature phone.

Voice callback advisory (IVR).

Telegram / WhatsApp photo diagnosis.

Full platform → about.agro-futures.com    View vertical →

The Rail Is Proven

Built on proven rails.

Proven at Scale
2.3B
Mobile money accounts globally, $2T transacted last year, on the same signal layer (GSMA State of the Industry 2026). Twitter ran on SMS shortcodes before smartphones existed. M-Pesa proved the same rail moves money. This is the most proven delivery layer on earth, and it runs on the signal that works when everything else fails.
The Precedent
2006
Twitter
Built as an SMS product first. 140 characters was a carrier limit, not a design choice.
2007
M-Pesa launches
A full mobile-money system on USSD menus, no smartphone or app required. Kenya's national payment rail before smartphones were common there.
2025
$2T moved globally
Mobile money processes roughly $2 trillion a year worldwide. The USSD gesture scaled from one country's payment habit to a global financial rail.
NOW
A different payload, same rail
40 million Kenyans already perform this exact USSD gesture every month. From The Ground Up puts a crop advisory, a compliance record, or a knowledge lookup on the gesture instead of a payment.
The Market

Payment verification is live across 21 African countries via PawaPay, covering an estimated 85% of continental mobile money volume. USSD and voice delivery run via Africa's Talking across overlapping coverage. The worker-facing layer is live in Kenya; deployment to additional countries is carrier registration, not new architecture.

$1.4tn
Transaction value, +26% p/a
341m
Active 30-day accounts, +19% p/a
85%
Of all African mobile money, 20+ countries

Source: PawaPay, pawapay.io (accessed Aug 2026). Figures are PawaPay's own published network stats, cited as evidence of the payment-rail scale From The Ground Up's compliance and advisory layer runs on top of, not From The Ground Up's own metrics.

Benin · Burkina Faso · Cameroon · Congo-Brazzaville · DRC · Ghana · Guinea · Guinea-Bissau · Ivory Coast · Kenya · Lesotho · Malawi · Mali · Nigeria · Rwanda · Senegal · Sierra Leone · South Africa · Tanzania · Uganda · Zambia

The senior companionship gateway, an architectural branch of the same rail, runs on Twilio in the United States, proving the infrastructure generalises across carriers, continents, and populations.

Full technical breakdown → rail.html

The Library

A 21st-century reference, spoken in your language.

Dholuo Wikipedia has about 800 articles. English has 6.8 million. Content exists; it doesn't exist in the language the reader speaks. The same rail that delivers a crop advisory already carries a glossary lookup, on USSD and WhatsApp/Telegram (/define), with a voice callback for anyone who can't read. Open source, live in AgroFutures today; the standalone general-knowledge release is next. View vertical →

Proof of Concept

One rail. Built from the ground up.

Each vertical below runs on the same delivery layer: USSD, voice, SMS, WhatsApp, Telegram, one backend. The same shortcode that delivers a crop advisory in Dholuo logs a labour shift for EU compliance and accepts a silent abuse report from any borrowed phone. Each vertical is labelled by the stage it is actually at. We don't round any of them up.

East Africa · EU Compliance EU Compliance Portal (DPP)
Live

A USSD labour-logging, wage-verification, and origin-traceability system for East African cooperatives, feeding a Digital Product Passport for EU buyers and auditors: CSDDD labour due-diligence, EUDR deforestation screening, and ESPR/DPP product traceability on one ledger.

View vertical →
East Africa · Agriculture AgroFutures
Live

Two tools on one rail: a farmer's per-plot advisory over 2G USSD, voice, WhatsApp, and Telegram (photo diagnosis included), and a Regional Portfolio Manager for the officer covering 1,000+ farmers, with pest-alert clustering, 4-language USSD registration, and human-gated advisory broadcast built on the same 35+ data points per farm.

View vertical →
East Africa · Logistics Cold Chain Intelligence
Pilot · In discussion

A routing and chain-of-custody layer for cold-storage logistics: fuel-optimised collector routing, USSD-witnessed dual-custody handoffs at the farm gate, and a harvest ledger that builds smallholders' first bankable credit history. In active pilot discussions with an East African cold-storage operator, not yet signed, not yet deployed.

View vertical →
United States · Social Care Senior Companionship
Open source · Concept

SMS-based AI companionship for isolated older adults without a smartphone, a direct architectural branch of the AgroFutures gateway, proving the rail generalises across unrelated populations and continents. Today it is an open-source reference implementation with no deployed users; we're seeking a US partner.

View vertical →
East Africa · Knowledge Access The Library
Live · Open source

A glossary lookup on the same rail used for farm advisory, live today on USSD and WhatsApp/Telegram (/define <term>): plain-language definitions with a voice callback for anyone who can't read. Designed around the same Gate 04 problem: Dholuo Wikipedia has ~800 articles, English has 6.8 million. A general-knowledge version, open to anyone with a phone, is next.

View vertical →
B2B · Infrastructure Custom Rail
Available

We build the delivery layer for other organisations that need to reach a population across the full device spectrum, from a $15 phone on 2G to a smartphone on WhatsApp. One backend, every channel, any country. You bring the use case.

View vertical →
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 ↗

knowledge-rail-2g

USSD/WhatsApp glossary lookup: Wikipedia extract + language fallback + voice callback. MIT.

GitHub ↗
Contact

Work with us.

We work with researchers, NGOs, development agencies, funders, EU buyers and compliance teams, cold-chain operators, and organizations building delivery infrastructure for populations no app can reach.

For partnerships, grant co-applications, custom rail builds, or technical collaboration:

[email protected]