South African AI research startup Yazi has secured its first institutional round, with the company now valued at ZAR30 million, or about $1.8 million, on a pre-money basis.

South African AI research startup Yazi has secured its first institutional round, with the company now valued at ZAR30 million, or about $1.8 million, on a pre-money basis.

The raise gives the business fresh backing as it tries to scale a model that uses WhatsApp as the main interface for research collection and analysis.

Yazi helps organisations run surveys, interviews, and broader research projects through WhatsApp, using AI to carry out conversations with participants and then process the responses.

That approach removes a common friction point in research work, since users do not need to download a separate app or log into a web platform before taking part.

For many respondents, especially in markets where WhatsApp is already part of daily communication, that simplicity can improve participation and response quality.

The startup already counts Old Mutual, Pick n Pay, Discovery, Capitec and Ipsos among its clients, which gives it a strong early signal that the product solves a real enterprise problem.

In research-heavy industries, the challenge is often not collecting opinions but collecting them quickly, at scale, and in a format that can be acted on. Yazi is trying to collapse that workflow into a channel people already know.

The undisclosed round was led by 3 Capital Ventures, the early-stage investment firm tied to Allan Gray. The capital will go into product development and international expansion, with the company planning automated voice interviews over WhatsApp as its next major feature.

It also wants to grow its pool of research participants across African markets, a move that could make the platform more useful for cross-border studies and consumer insight work.

From my experience covering African startups, Yazi sits in one of the more practical corners of applied AI. It is not chasing broad consumer attention or abstract model hype.

It is using AI to solve a specific workflow problem in enterprise research, where speed, accessibility, and respondent engagement matter.

The real test will be whether it can preserve research quality as it scales into voice, new geographies, and larger customer bases. If it does, the company could become a useful example of how AI can sit inside existing communication habits instead of asking users to change them.

This post was culled from Disrupt-Africa.

Don’t miss important articles during the week. Subscribe to techbuild weekly digest for updates