SHIFA Selected for the AI UniPod Bootcamp, Second Cohort β€” Powered by timbuktoo, UNDP Ethiopia & the AI Institute πŸš€

What we build

Two live AI products, one mission

AI for the frontline of Ethiopian healthcare β€” designed locally, in local languages, for resource-constrained health systems.

A computer circuit board with a brain graphic, representing AI-powered clinical decision support
Live product

SHIFA Clinical Intelligence Platform (SCIP)

An AI clinical decision-support tool that gives frontline health workers real-time, guideline-based answers at the point of care.

Selected for the African Impact Challenge 2026 Bootcamp (Health Entrepreneurship – Builder Track) and the AI UniPod Bootcamp, Second Cohort β€” powered by timbuktoo with UNDP Ethiopia, the Ethiopian Artificial Intelligence Institute, and Addis Ababa University.

SCIP gives frontline health workers instant, cited answers instead of a stack of paper protocols β€” grounded in 109 validated Ethiopian and WHO clinical guidelines, so every answer traces back to a real source rather than a guess.

It's built for the point of care: a health worker asks a clinical question in the moment, and SCIP returns a guideline-based answer, not a generic search result.

Live product

Degdeg

AI Climate-Health Early Warning

An AI early warning platform that turns open climate forecasts into anticipated health impacts and plain-language advisories for woreda health workers β€” so health offices can act before disaster becomes an emergency.

100

Woredas monitored

8

WHO & national protocol sources

~2,000

Indexed protocol passages

100%

Advisories human-approved

Degdeg closes three gaps that keep strong early-warning science from reaching the people who need to act: forecasts rarely reach woreda health offices in a form they can act on; a rainfall anomaly map doesn't tell a health officer to expect a cholera risk window in flood-hit kebeles; and even when warnings arrive, nothing converts them into a protocol-based preparedness checklist.

The platform runs a four-stage AI pipeline β€” Interpret, Anticipate, Advise, Approve. It reads 14-day climate forecasts and disaster bulletins to produce a graded hazard assessment, checks that assessment against 15 vetted climate-to-health impact rules, drafts advisories in Somali and English (WhatsApp message, radio script, official memo) grounded only in cited risks β€” and publishes nothing automatically. Every advisory is reviewed and approved by a health officer before it reaches a woreda.

First deployed across 100 woredas in the Somali Region, built to scale nationally β€” grounded in 8 WHO and national protocol sources with ~2,000 indexed protocol passages.