Mohammad Abu Sheikh: Pioneering Sovereign AI Innovation in the MENA Region

Driving AI Innovation in the MENA Region

Mohammad Abu Sheikh is revolutionizing artificial intelligence across the Middle East and North Africa by shifting the focus from passive consumption to sovereign innovation. As the CEO of CNTXT AI and the founder of a $10 million AI fund, he has successfully led three business exits and secured over a billion dollars in funding. His vision is to create an AI ecosystem deeply rooted in language, culture, and data sovereignty.

CNTXT AI's Mission and Vision

CNTXT AI specializes in digital transformation, offering cloud infrastructure, industrial software, and robotics solutions designed to modernize operations and unlock data-driven insights across the MENA region. The company was founded upon recognizing an abundance of underutilized data and addressing the lack of AI readiness caused by insufficient data preparation.

Overcoming Challenges in Building Arabic AI Data Resources

One of the biggest obstacles has been the scarcity of high-quality Arabic data online. Arabic content is seriously underrepresented, making up only 3–5% of all online content. To overcome this, CNTXT AI deployed data labelers, annotators, and data scientists to digitize and curate Arabic data, creating a foundation for culturally relevant AI models.

Emphasizing Cultural Relevance in AI Development

CNTXT AI embeds culture at every stage of AI development, from infrastructure to final products. Recognizing the diversity of dialects and cultural contexts within the Arabic language, they build localized products tailored for specific countries by working with annotators native to those regions.

Complementary Ventures: LocAI and SMPL AI Fund

LocAI serves as the application layer built upon CNTXT AI’s data and infrastructure, delivering real-world AI solutions. The SMPL AI Fund invests in early-stage startups to nurture the regional AI ecosystem by sharing tools and insights gained from CNTXT AI’s experience.

Developing Munsit: The Most Accurate Arabic Speech Recognition Model

Munsit was developed out of necessity to meet demand from government and private sectors. Unlike other models adapted from English technologies, Munsit was designed Arabic-first. It uses a weakly supervised learning approach, training on 30,000 hours of unlabeled Arabic speech to create a high-quality 15,000-hour dataset without relying on costly manual transcription.

Impact and Future of Arabic AI Innovation

Munsit’s performance surpasses models from major global players like OpenAI, Microsoft, and Meta, signaling the rise of regionally owned AI technologies. This success underscores the importance of data sovereignty and the ability to build world-class AI solutions by local talent for local needs.

Looking Ahead: Expanding Arabic Voice AI

CNTXT AI plans to release a new suite of Arabic-first AI solutions powered by Munsit and other models. This marks only the beginning of their journey to advance Arabic voice AI capabilities.

The Importance of Sovereign AI

Sovereign AI means full ownership and control over data, infrastructure, and AI models, which is critical for the Gulf and broader MENA region. Ensuring data sovereignty allows the region to own its technological future rather than relying on external parties.

Shaping the AI Ecosystem in the Middle East

CNTXT AI is committed to enabling AI readiness by helping organizations build data and AI strategies, then assisting in building, testing, deploying, and scaling AI solutions. They focus on refining unstructured data to unlock its potential, equating data to a new form of oil.

Advice for AI Startups in Emerging Markets

Mohammad advises entrepreneurs to start immediately, move quickly, learn from failures, and iterate continuously. The key to success lies in solving real problems and staying close to users’ needs, emphasizing relevance and adaptability in emerging markets.

For more information, readers are encouraged to visit CNTXT AI.