Voice AI for customer support is a rapidly growing area, yet developing products that sound natural and respond instantly proves considerably more challenging in certain markets. Specifically, most leading voice AI solutions were not initially conceived with the unique requirements of Africa and the Middle East in mind.
Addressing this critical gap, AethexAI, a startup established last year, has successfully secured $3 million in pre-seed funding. The round was spearheaded by 4DX Ventures, with additional participation from Enza Capital, Dorm Room Fund, Mojo Ventures, and Stanford GSB 26 Fund. Notable individual investors include Stanford faculty members, telecom executives, and AI researchers from Anthropic.
Diverging from reliance on existing orchestration tools such as Vapi and LiveKit, AethexAI strategically developed its own compact model and orchestration layer from the ground up. This bespoke approach was adopted to effectively manage the diverse localized dialects of English, French, and Arabic prevalent across its target markets, a decision fundamentally influenced by the specific operational demands of the region.
Concurrently, the company is rolling out its platform, inviting enterprises to explore its technology and subscribe to its services. It is also providing APIs and SDKs to enable developers to experiment with its proprietary models.
The startup was co-founded by Mariama Diallo and Ayooluwa Odemuyiwa. CEO Diallo previously worked at Goldman Sachs before joining YC-backed ModelML in a product and growth capacity. CTO Odemuyiwa, a Caltech alumnus, gained experience at Meta and was enrolled at Stanford Business School prior to co-founding AethexAI. Their shared ambition was to create solutions tailored for emerging markets, prompting their search for relevant opportunities.
Businesses globally are actively pursuing AI tools to automate operational segments; however, success is not always guaranteed. In Egypt, for instance, the founders discovered that a call center automated a significant portion of its calls but subsequently reverted the system due to unsatisfactory outcomes. Furthermore, several support centers across Africa communicated the persistent challenge of sourcing and employing engineers to automate calls cost-effectively.
"The latency and jitter that we saw on automated calls in this region were outrageous," Odemuyiwa explained to TechCrunch, elaborating on the rationale behind developing the company's own models and orchestration layer. He added, "If we had become orchestrators, we might have had to use large models that were hosted outside the region, resulting in higher latency. We realized that in order for this to work, we have to use very small models and cut latency at every step."
While typical AI labs invest millions in training and data acquisition for their advanced models, AethexAI devised an innovative solution for both challenges. Instead of pursuing the largest possible models, the company determined that smaller models were sufficient to effectively address the latency issue while preserving accuracy. This led to the development of its Kora series, featuring parameters ranging from 300 million to 1.7 billion – a deliberate fraction of the size of conventional Large Language Models (LLMs).
To train these specialized models, the startup leveraged anonymized recordings from a call center partner. Additionally, it dispatched hard drives to radio stations across Africa to gather a broader spectrum of audio data. To maintain cost efficiency, AethexAI established a contributor network of university students tasked with annotating data and accurately pronouncing local names. As a direct result of these efforts, the startup reports currently managing over 17,000 calls daily.
From a business perspective, the company is committed to guiding clients new to voice AI through the implementation process. This includes providing onsite demonstrations and workshops to assist them in identifying the most suitable use cases for automation.
"We always tell customers that we cannot be everything for everybody right now. We’re small," Diallo stated, emphasizing their focused strategy. She continued, "When we start talking to a company, we ask them to pick one use case that is the most important to them to start [with]." While the startup is open to collaborating across all industries, a significant portion of its current applications involves calls for debt collection, customer activation, or Know Your Customer (KYC) verification – the standard identity-checking protocol used by banks and telecoms. To serve local markets, the company is hiring forward-deployed engineers on a contract basis and forging channel partnerships with telecom providers to manage telephony for voice AI calls, underscoring that simple "plug-and-play" solutions are inadequate for the region.
Walter Badoo, co-founder and managing partner of 4DX Ventures, asserts that the markets in Africa and the Middle East fundamentally differ from those for which most voice AI companies were originally designed.
He elaborated, "Enterprises in Africa and the Middle East process roughly three times the call volume of their Western counterparts, as voice is still the dominant channel for customer interaction." Badoo further noted, "Incumbent systems were built for Western markets characterized by high-end GPU infrastructure, standard English and European speech environments, and enterprise workflows common in the US and Europe. That creates real gaps when enterprises need systems that handle dialects, code-switching, and informal speech patterns, and that work within their existing telephony infrastructure and their actual price points."
In essence, while companies such as ElevenLabs, Deepgram, Sierra, and Cognigy are rapidly expanding globally, the markets they were initially built for and the new markets they are entering are often distinct. Startups like AethexAI are capitalizing on the belief that these unaddressed gaps — including models specialized in local dialects, robust on-the-ground partnerships, and infrastructure specifically developed for the region — represent a significant market opening that larger industry players lack both the incentive and the architectural framework to adequately address.
The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.