Founders Depart Goldman and Meta to Develop Underserved Voice AI Markets
Image Credits:aethexai
The Transformation of Customer Support through Voice AI
Customer support and service are currently among the leading sectors poised for transformation through voice AI technology. However, creating a product that resembles human interaction and responds seamlessly is significantly more challenging in certain regions. Notably, many of the major voice AI platforms were not designed with the unique requirements of Africa and the Middle East in mind.
The Emergence of AethexAI
To address these challenges, AethexAI—a startup founded last year—aims to fill the existing gap in the market. Recently, the company successfully secured $3 million in pre-seed funding led by 4DX Ventures, with additional support from Enza Capital, Dorm Room Fund, Mojo Ventures, and Stanford GSB 26 Fund. Among its investors are Stanford faculty, telecom industry leaders, and AI specialists from Anthropic, demonstrating a robust backing for its vision.
Custom Solutions for Local Dialects
Instead of relying on standard orchestration tools like Vapi and LiveKit, AethexAI has developed its own compact model and orchestration layer tailored for the diverse dialects of English, French, and Arabic dominant in its target regions. This strategic choice was influenced by the distinct operational needs present in these areas, showcasing the company’s commitment to regional specificity.
In tandem with launching its platform, AethexAI aims to provide enterprises the opportunity to test its technology, register for services, and access APIs and SDKs for developers to experiment with its models.
Founders with a Vision for Emerging Markets
AethexAI was co-founded by Mariama Diallo and Ayooluwa Odemuyiwa. Diallo previously worked at Goldman Sachs and joined YC-backed ModelML for product and growth initiatives. Odemuyiwa, a Caltech graduate, has experience from Meta and is currently enrolled at Stanford Business School. Together, they recognized the potential to deliver innovative solutions specifically designed for emerging markets.
Real-World Challenges in AI Implementation
While businesses globally are eager to adopt AI tools for operational automation, success is not assured. For instance, a call center in Egypt automating its operations ultimately reversed the changes due to unsatisfactory outcomes. Founders of AethexAI encountered feedback from several support centers in Africa stating that hiring engineers for cost-effective automation was a consistent challenge.
Odemuyiwa shared, “The latency and jitter we observed on automated calls in this region were unacceptable. Had we opted for standard orchestration, we would have had to utilize larger models hosted outside the region, which would have resulted in more significant latency. Thus, we realized we had to focus on smaller models and minimize latency at every turn.”
Innovative Model Development: The Kora Series
While many AI laboratories spend substantial amounts on training their models and acquiring data, AethexAI found a smarter approach. It decided to focus on smaller models, acknowledging that these could adequately manage latency without sacrificing accuracy. The company’s Kora series ranges from 300 million to 1.7 billion parameters, a fraction compared to larger language models, which aligns perfectly with its objectives.
To train these innovative models, AethexAI utilized anonymized recordings from a partner call center. Additionally, it sent hard drives to radio stations throughout Africa to gather further audio data. To keep costs manageable, the startup also engaged a contributor network composed of university students tasked with data annotation and pronouncing local names. These efforts have resulted in AethexAI managing over 17,000 calls daily.
Walking Clients Through the AI Ecosystem
On the business front, AethexAI is careful to guide clients unfamiliar with voice AI technology through the onboarding process. It offers on-site demonstrations and workshops aimed at helping enterprises identify the most relevant use cases for automation.
Diallo explained the company’s focused approach: “We always inform our customers that we can’t cater to every need at this moment. Our resources are limited. When we engage with a company, we encourage them to select one primary use case to start.”
Target Industries and Use Cases
While AethexAI is open to opportunities across multiple sectors, its current focus includes applications in debt collection, customer activation, and KYC (Know Your Customer) processes—essential for identity verification in banking and telecommunications. To optimize its services, the company is hiring engineers on a contract basis to adapt solutions for local markets and is establishing partnerships with telecommunications providers for voice AI call management. The traditional plug-and-play approach, they assert, is not feasible within their operational landscape.
Understanding the Unique Market Dynamics
Walter Baddoo, co-founder and managing partner of 4DX Ventures, emphasizes that the market dynamics in Africa and the Middle East differ significantly from those for which most voice AI systems were originally created. “Enterprises in Africa and the Middle East typically manage about three times the call volume of their Western counterparts, as voice remains the predominant method of customer interaction,” he notes.
He further elaborates, stating that established systems were largely designed for Western markets equipped with advanced GPU infrastructure, predominately standard English, and specific enterprise workflows typical of the U.S. and Europe. This results in clear gaps for enterprises seeking systems that can accommodate local dialects, code-switching, and informal speech patterns while integrating smoothly with existing telephonic setups and budgetary constraints.
Seizing Market Opportunities
In a landscape where companies like ElevenLabs, Deepgram, Sierra, and Cognigy are rapidly expanding, AethexAI is strategically positioning itself to capitalize on market gaps—particularly in creating models fine-tuned for local dialects, fostering on-ground partnerships, and building infrastructure aligned with regional needs. Essentially, AethexAI is betting that these localized efforts represent a significant market opportunity that larger companies may lack the incentive—or the capability—to approach effectively.
By creating solutions tailored to customer interactions in Africa and the Middle East, AethexAI is not just stepping into the voice AI space; it’s redefining the possibilities of customer support and service, ensuring that emerging markets are no longer an afterthought in the global tech narrative.
Thanks for reading. Please let us know your thoughts and ideas in the comment section down below.
Source link
#founders #left #Goldman #Meta #build #voice #markets #overlooked
About The Author
