Botlhale AI: Pioneering Inclusive African Language Tech for AI Growth

The artificial intelligence (AI) sector in South Africa - and across the African continent - holds immense potential. With a growing number of startups leveraging AI to tackle the continent’s unique challenges - one key issue stands out: the integration of African languages into natural language processing (NLP) systems.

Africa is home to an estimated 1500 to 3000 languages, according to UNESCO. With the growth of new technologies such as AI, there is an urgent need for solutions built by Africans, for Africans - and that operate in the languages people actually speak. Tech that integrates African languages is a part of a growing trend of inclusive innovation emerging from the continent. Leading this charge is Botlhale AI.

A South African Startup Championing Responsible, Localised AI

Founded in 2020, Botlhale AI is a South Africa-based startup that develops AI-powered NLP tools, enabling businesses to connect with customers in languages they understand and trust. In an interview with Tech Africa, the company emphasised that responsible AI is a cornerstone of its development ethos. Botlhale ensures that all user data is anonymised and securely stored, with clear consent mechanisms in place.

To tackle one of the most pressing challenges in AI — bias — the company relies on diverse and representative datasets. Community annotators are actively involved in the data labelling process to capture authentic linguistic and cultural nuances.

“We deliberately seek linguistic diversity in our data sourcing. Our community annotation model involves speakers from various regions and dialect groups. We avoid overfitting by balancing datasets, validating model outputs across dialectal variations, and continuously expanding language coverage beyond just the most spoken ones. Minority dialect inclusion is not an afterthought — it’s part of our commitment to broad, ethical representation,” says Thapelo Nthite, CEO of Botlhale AI.

NLP Tools for Business: Bua, Vela and Speech APIs

Botlhale AI has developed three main products: Bua, Vela, and a suite of speech and language APIs.

Bua is a collection of NLP tools that allow service providers to interact with customers through digital platforms in various African languages. Vela is a generative AI-powered call centre analytics API, that detects and captures spoken languages in phone calls. The company’s speech APIs support a range of African languages including isiZulu, isiXhosa, Kinyarwanda, Swahili, Tshivenda, and Sepedi, among others.

These tools are already in use across various industries. One example is the DSTV WhatsApp chatbot, which saw improved user engagement and satisfaction following its integration with Botlhale’s technology. Another is a collaboration with Chenosis, an accelerator platform that enables companies to communicate with customers in their native languages via conversational AI. 

“We look for collaborators who share our vision of digital inclusion and who can help us scale infrastructure, access diverse datasets, or co-develop use-case-specific solutions. Whether it's telcos, fintechs, or customer experience platforms, we value partners that enhance our capabilities and share our commitment to solving real problems in African contexts,” explains Nthite. 

Balancing Global AI Trends with African Realities

Botlhale AI strikes a rare balance between global relevance and local impact. By building globally competitive infrastructure - such as speech recognition, NLP, and voice tech - and adapting it for African languages, the company proves that scalability and local relevance are not mutually exclusive. 

“African communities are rich in linguistic, cultural, and contextual diversity. While many global AI solutions tend to default to English or other dominant languages, we build with a deep understanding that language is not just a tool of communication - it's a carrier of identity, trust, and inclusion,” says Nthite. 

All of Botlhale’s tools are built with this duality in mind: they integrate easily into global systems but remain sensitive to the day-to-day realities of African businesses and users.

“By grounding innovation in local realities, we’re not just adapting global tech - we’re contributing to it, showing the world that African languages and contexts deserve, and can power, world-class AI,” Nthite adds. 

Community-Centric AI: Co-Creating With Real Users

Central to Botlhale AI’s strategy is its community-based data annotation platform, which involves native speakers and underrepresented groups in shaping how the models understand and generate language.

“By co-creating with real users - from call centre agents to customers - we ensure our tools are grounded in real-world needs and contexts. Our model improvements are often guided by feedback from the very people using them,” he says.

Driving Africa’s AI Future Through Language and Inclusion

With Africa set to have the world’s youngest population by 2030, inclusive technology is not just ideal - it’s essential. Botlhale AI exemplifies how to build scalable solutions for Africa that also compete on the world stage. 

Looking ahead, the company sees itself as more than a product company — it wants to be an ecosystem enabler for Africa’s AI future.

“We want to drive the infrastructure and community needed for Africa to own its AI future,” Nthite concludes. “We’ll continue to champion linguistic inclusion, ensure responsible AI practices, and foster a tech ecosystem where African languages are not just supported — they lead.”