Intella, a speech intelligence company focused on Arabic dialects and headquartered in Riyadh, has raised US$12.5 million in a series A round led by Prosus, with participation from 500 Global, Wa’ed Ventures, Hala Ventures, Idrisi Ventures, and HearstLab.

The company plans to use the funding to expand its enterprise offerings, grow its presence in Egypt and Saudi Arabia, and invest further in R&D.

Founded in 2021, Intella develops AI-powered transcription, analytics, and customer engagement tools for businesses in the Middle East and North Africa.

Intella recently launched Ziila, a conversational AI agent, which was first deployed with ecommerce platform Jumia to support voice-ordering features.

Intella has more than doubled its revenue in 2024 and is on track for 7x growth in 2025.

With this round, Intella’s total funding raised stands at over US$16.9 million.

🔗 Source: Intella

🧠 Food for thought

1️⃣ Arabic AI presents massive market opportunity hampered by technical complexity

The Arabic-speaking market represents substantial untapped potential, with Arabic ranking as the fifth most spoken language globally and over 7,500 companies operating across the MENA region 1.

However, this market has been historically underserved by AI technology due to Arabic’s unique linguistic challenges.

Research shows that Arabic has a fundamentally different and more complex morphological structure compared to European and East Asian languages, creating significant obstacles for natural language processing tasks 2.

The complexity extends beyond grammar to include issues like word ambiguity and broken plurals, where plural forms don’t resemble their singular counterparts, making automated processing particularly difficult 2.

This technical barrier explains why Prosus investor Robin Voogd noted that “Arabic AI models have historically underperformed” and why the market remains largely addressable despite its size 1.

2️⃣ Dialect-specific AI development proves essential for market penetration

Intella’s approach of building models for over 25 Arabic dialects addresses a fundamental requirement that generic AI solutions cannot meet.