When you call a customer care service line, you’re likely to hear from a representative that “this call is recorded for quality purposes.” The promise of that message, ensuring quality, is what South Africa’s Botlhale AI is betting on, particularly for conversations that happen in African languages.
Founded in 2019, Botlhale AI provides businesses with conversational AI tools and multilingual call center analytics. Their core offering is a suite of natural language processing (NLP) tools—including speech-to-text, text-to-speech, and language understanding models—that help call centres to transcribe, analyse, and extract insights from customer interactions.
South Africa has 12 official languages, yet most call centre technologies and quality assurance systems have historically prioritised English. For millions of South Africans, English is not the language in which they feel most comfortable expressing themselves, especially when dealing with complex issues like insurance claims or service complaints.
While calls are routinely recorded for “quality purposes,” Thapelo Nthite, cofounder of Botlhale AI, told TechCabal that up to 70% of call centre conversations happen in languages other than English, but most automated quality assurance tools can only process English calls. This means that for the majority of interactions where customers switch between isiZulu, isiXhosa, Sesotho, Setswana, or other local languages, companies rely on manual review, if any review happens at all.
“That creates serious risk as critical information can be missed, compliance failures can go undetected, and vulnerable customers may not get the support they need,” Nthite said. “For example, in financial services, a missed medical disclosure in isiXhosa during an insurance sales call could result in a denied claim later, potentially exposing companies to regulator fines and brand damage.”
Business model
Botlhale operates on a Software-as-a-Service (SaaS) model. Clients, typically enterprise call centres, either pay per license (number of agents using the system) or based on consumption (e.g., minutes of audio processed).
“It’s a recurring SaaS fee, as some clients want to monitor a percentage of their total volume—say 20% of 500,000 minutes—and we price accordingly,” Nthite said.
According to Nthite, over 85% of consumers in Africa want to interact with brands in their native languages. You will often hear five languages on a single call, Nthite said. Additionally, audio is often low-quality and conversations tend to be fast-paced. Yet, many call centres still automate only English language calls. “If you are serving the South African market, about 70% of calls are not in English. That is a huge blind spot,” he says.
Currently Botlhale AI translates 11 South African languages and others like Setswana and Sesotho spoken in its neighbouring countries making it easy to enter those markets. Some of Botlhale AI’s clients in South Africa already operate across African markets like Botswana, Ghana, Kenya, and Nigeria which represent major markets ripe for expansion, especially as mobile penetration deepens and financial services grow. Botlhale AI’s expansion strategy is to follow them into these territories. “Our customers are leading us. If they say they are launching in Ghana, we build support for Ghanaian languages,” Nthite said.
Big bets
Botlhale AI’s big bet is that African languages will soon be supported at a very high level of quality by language technologies.
Over the next five years, Nthite highlighted that AI adoption in Africa will see significant growth especially across key language processing tasks like translation, transcription, text-to-speech, and language understanding. Access to high-quality tools will become more widespread, but right now, support for these languages is still patchy. “We have heavily invested in speech technology in the customer service space because we see a huge potential. For instance, call centres are already beginning to automate or augment parts of their operations, and we are tailoring our products to solve today’s challenges while anticipating future needs,” Nthite noted.
Right now, Botlhale AI is focused on augmenting quality assurance (QA) processes, but also “thinking ahead about what the future of call centres will look like in a multilingual, AI-driven world,” according to Nthite who argues that African languages could also be preserved through digitisation and development of language models.
The next biggest release
At the backend, Botlhale claims it is refining its transcription models and looking to release them soon through APIs making them accessible to third-party developers
“We want people to build with our tech, whether it is apps for learning languages or education tools, that’s part of how we scale impact,” Nthite said, adding that opening access to the company’s technology will allow its impact be felt in other areas from healthcare to education and law.
————
Mark your calendars! Moonshot by TechCabal is back in Lagos on October 15–16! Join Africa’s top founders, creatives & tech leaders for 2 days of keynotes, mixers & future-forward ideas. Early bird tickets now 20% off—don’t snooze! moonshot.techcabal.com
