When language becomes a gateway, not a barrier
In Dakar, the promise is simple yet decisive: to speak to artificial intelligence in the language people use every day. For millions of Senegalese, Wolof is the language of everyday interaction, while Pulaar remains a language of connection in several regions of the country and beyond. When digital technology ignores these practices, it leaves part of the public on the sidelines. This is precisely the barrier that AWA, a conversational assistant developed in Senegal, aims to remove.
The stakes go beyond ease of use. French is Senegal’s official language, but the Constitution also recognises several national languages, including Wolof, Pulaar, Serer, Jola, Mandinka and Soninke. This linguistic reality shapes education, commerce, local administration and mobile use. At the same time, UNESCO notes that the country is making progress in promoting national languages with tools such as Sentermino, designed to structure knowledge in Wolof, Pulaar and Serer. The constitutional recognition of national languages is therefore not a symbolic detail: it opens up a practical space for innovation.
AI designed from the ground up in Senegal
AWA is the creation of Senegalese start-up Andakia, founded in Dakar by two entrepreneurs, Alioune Badara Mbengue and Pape Gorgui Ndoye. Their aim is to bring local practices into AI tools rather than make users adapt to interfaces designed elsewhere. The idea is ambitious: to turn the assistant into a gateway to services, information and everyday tasks, first in Wolof, then in Pulaar and other African languages.
This approach addresses a problem familiar to researchers and digital-sector players. A study by AI for Development estimates that African languages as a whole account for only about 0.02% of global internet content. This directly affects the quality of AI models, which learn from the data available online. The ITU also notes that access to inclusive AI in Africa depends on data, infrastructure and services capable of operating in local languages. The under-representation of African languages in digital data is therefore not just a statistic: it is a structural barrier to useful innovation.
The choice of Wolof is strategic. It is the most visible language in Senegalese public life, and one that can quickly reach a large number of users, including people who are less comfortable with written text. Pulaar, meanwhile, extends the project’s reach to a linguistic community found in Senegal as well as several countries across West Africa and the Sahel. By targeting these languages, Andakia is doing more than translating words. It is trying to adapt technology to the way people actually ask for help, look for a service or express a need.
What changes when AI speaks local languages
The first impact is social. A voice assistant in Wolof or Pulaar can help a shopkeeper, pupil, parent or user unfamiliar with keyboards interact more easily with digital technology. In a country where access to smartphones has grown faster than familiarity with complex interfaces, voice becomes a bridge. This matters for education, healthcare, local commerce and public services.
The second impact is economic. Technology adapted to local languages opens up an under-served market: call centres, information services, educational platforms, customer support, training tools, voice commands and information searches. It is also a way to retain more value on the continent. Rather than relying exclusively on imported tools, Senegalese companies can create products tailored to local needs, then export them to other African markets.
The third impact is political, in the broad sense. Senegal has already made digital sovereignty a subject of public debate, particularly in relation to education, data and infrastructure. UNESCO says the country is among the African pilots integrating digital skills and AI into education systems. In this context, AI that understands national languages fits a coherent approach: bringing technology closer to the population instead of reserving innovation for those who know the web’s dominant conventions. UNESCO’s work on AI skills and local languages in Africa confirms that this is not a marginal trend.
Reaching beyond Dakar
AWA is aimed first and foremost at Senegal, but its appeal extends much further. Pulaar is spoken in several West African countries, and the very principle of AI rooted in national languages could inspire other ECOWAS markets. In a region where digital use is growing rapidly, the question is no longer simply how to connect people. It is also which languages they can actually use to access services.
This issue also highlights an obvious point often overlooked in debates about artificial intelligence: linguistic diversity is not a cultural add-on. It is a condition of access. African players investing in this area, in Dakar and elsewhere, are not simply seeking to imitate the world’s major platforms. They are building tools grounded in local realities, everyday practices and living languages. An important part of Africa’s AI future will be shaped here.



