Kivuli Index Lab studies how Kenyan businesses are represented in AI answers across sectors, counties, languages and business forms. Its work is practical, slow enough to be checked, and focused on the gap between local business reality and machine-readable evidence.
A Nairobi tour operator showed up again and again in the lab’s founding comparison. A coastal operator with a working site, reviews and licence references did not. That plain comparison became the first question behind Kivuli Index Lab: why does one Kenyan business travel through AI answers while another similar business stays in the shade? The answer did not look like a single source problem. It looked like sector visibility, regional gravity, language choice and business form all pulling on the result at once.
The group formed around that knot. Amara Njoroge brought benchmark framing from small-business support research. Silas Mwenda shaped prompt runs and comparison sheets. Nia Otieno watched the English-Swahili shifts that can change how a sector is described. Baraka Kilonzo kept attention on business forms that are easy to miss in tidy datasets: jua kali enterprises, cooperatives, SACCOs and mobile-first sellers. Together, they built a method that records answer states before making claims about causes.
The lab’s position is deliberately narrow and useful. AI visibility in Kenya should be read as infrastructure: uneven, improvable and measurable when the method is careful enough. That means looking past dramatic rankings and asking plainer questions. Who is named? Who is absent? Which county gets treated as the default? Which business is compressed into a generic label? Those questions are not glamorous. They are the screws holding the benchmark together.
Sector-and-region visibility patterns across Kenyan business categories
She previously built practical research briefs for small-business support programs and regional market guides. Her work keeps the lab’s findings usable outside a research room.
Repeatable query sets across ChatGPT, Gemini, Perplexity, Google AI Overviews and Copilot
He previously worked on search-quality reviews, content audits and structured comparison sheets for service businesses. He is the person most likely to notice when a run cannot be reconstructed cleanly.
English and Swahili wording differences in Kenyan sector representation
She previously edited bilingual business explainers and public-facing service descriptions for local markets. Her focus is where translation changes the business meaning, not just the wording.
Gaps between registered firms, jua kali enterprises, cooperatives, SACCOs and mobile-first sellers
He previously prepared field notes on enterprise categories, trade clusters and county-level business information. His work keeps informal and hybrid enterprise forms visible in the method.
For topic suggestions, benchmark questions or careful critique, the contact form is the right door.
Contact the lab