Logo

About Tonative

Data Curation for Under-Resourced Languages

Who We Are

Tonative is an African AI data company. Our goal is to ethically provide high-quality datasets that teach AI to understand Africans better. This means the data we curate reflects how Africans actually speak a language, and how they interact with industry domains in ways unique to their country or region.

By closing this data gap, we enable AI models to become more reliable and serve a wider audience than it currently does. Our work is successful because we collaborate with a trained network of linguists and industry experts who live and work across the continent

Founder's narrative

The rapid growth of AI research and innovation across Africa has created unprecedented opportunities. As global investment, research partnerships, and technological interest in African markets accelerate, the demand for language models that understand local linguistic and contextual diversity has increased.

However, the continent's data needs have outgrown what grant-funded and volunteer-driven efforts can sustain. Meeting this demand requires a scalable, commercial-grade data infrastructure.

Tonative was built to be the data layer that enables African AI innovation by combining trained domain experts with AI-assisted curation workflows to preserve and scale Africa’s cultural and contextual knowledge

Why We Exist

Most AI systems today are trained on data from a small number of dominant languages. This creates systemic exclusion for communities whose languages and cultural contexts are missing or misrepresented.

Credentials

Published at NeurIPS 2025 Conference

Published at NeurIPS 2025 Conference

Human-AI collaborative data curation methodology.

HealthBench Extended

HealthBench Extended

Localised OpenAI HealthBench dataset to 6 Nigerian languages with medical practitioners rewriting scoring rubrics to match the Nigerian healthcare system.

MLC data validation partner

MLC data validation partner

Validated custom dataset for research purposes.

Featured in TechCabal YPIT

Featured in TechCabal YPIT

Recognized in Africa's AI ecosystem as a dataset infrastructure provider.

How We Work

Every dataset passes through a multi-stage pipeline. Our trained curators produce initial data, the quality assurance layer validates the data, the language or domain leads resolve inter-annotator disagreement before the final data is licensed. This process applies equally to both the custom and open-source datasets.

Our Vision

We envision a future where African languages can significantly participate in AI systems, and its communities have agency over how their languages and experiences are represented in technology.

Meet Our Team

Sharon Ibejih

Sharon Ibejih

Founder

Cynthia Amol

Cynthia Amol

Co-Founder & Head of Data

Alfred Kondoro

Alfred Kondoro

Head of Research

Damilare Keshinro

Damilare Keshinro

Head of AI Translations

Chinenye Anikwenze

Chinenye Anikwenze

Engineering Lead

Joy Naomi Olusanya

Joy Naomi Olusanya

Training Manager