The Kilwa Intelligence Platform
Investment decisions, not data dumps.
Six integrated modules that turn complex, multilingual, data-scarce market inputs into two answers: where to invest, and when.
The platform
Four surfaces, one answer.
Where to invest, when to invest, what the market is saying, and a way to ask. Each view reads from the same scored inputs, so nothing has to be reconciled afterwards.
ISI · Where to invest
Every market, sized by what it is worth.
All 54 markets at once — box area by GDP, colour by performance, an ISI marker on every market above 70. Filter by sector and the continent re-scores.
METI · When to invest
A posture, a window, and the drivers behind both.
Every market carries an OPTIMAL, WATCH or AVOID posture with an entry window, decomposed into the six factors that produced it and the dated catalysts that could move it.
Sentiment Pulse · Market mood
Read in the language it was published in.
Sentiment is extracted from French, Arabic and Swahili sources directly rather than translated into English first, because translation is where policy nuance is lost.
Zawadi · Ask the research
Every figure traces back to a scored input.
Ask about any market, sector or score Kilwa publishes. Zawadi answers from the published research and shows its sources — it does not know your holdings and does not take them as an input.
Illustrative platform output. Scores, performance and headlines shown are sample values, not live data, and news attributions are shown by source type rather than by publisher. The platform launches separately — see the methodology for how every score is built.
Modules
Each one feeds the same answer.
Nothing here is a separate dashboard you have to reconcile. Sentiment feeds the scores, the scores feed the narrative, and the explanation travels with all of it.
How it works
From raw multilingual data to a committee-ready thesis.
Ingest multilingual data
Macroeconomic series, policy announcements and regional news across 54 African markets, drawn from 416 registered sources in their original languages.
Model in parallel
Structured time-series and unstructured sentiment enter a hybrid model stack rather than being reconciled after the fact.
Score suitability and timing
ISI returns a sector-aware score out of 100. METI returns an ENTER, HOLD or WATCH posture with a 1–6 month window.
Explain and answer
Every score decomposes into its drivers, and Zawadi turns the result into a thesis you can put in front of a committee.
Built for institutional process
The score has to survive a risk committee.
Accuracy is necessary and not sufficient. An institutional user needs to know what drove a number, where the input came from, and what would make it wrong.
90
indicators behind every country score
85
languages in the registry
AES-256
encryption in transit and at rest
SOC 2
Type II controls, aligned
Get started
See the platform on the markets you follow.
A working session with your team, run across Kilwa’s covered markets.