Saudi Artificial Intelligence startup launches Arabic LLM

Misraj Artificial Intelligence unveiled Kawn, an Arabic large language model, at AWS re:Invent and launched Workforces, a platform for creating and managing Artificial Intelligence agents for enterprises and public institutions.

Saudi Arabia-based lab Misraj Artificial Intelligence revealed Kawn, a new Arabic large language model, at AWS re:Invent. The release includes a suite of capabilities and services aimed at supporting Arabic communication across industries. Key components named in the launch include Mutarjim, a bidirectional Arabic-English translation model, and Lahjawi, which the company describes as the first translation model that works across 15 Arabic dialects. The announcement also highlighted Kawn Document OCR, a vision-language model designed to identify and extract Arabic text from documents for downstream processing.

Safwan AlModhayan, CEO of Misraj, framed Kawn as an effort to provide a “dependable and enterprise-grade” Arabic Artificial Intelligence foundation. He told AI Business that the model was created to address what Misraj sees as a major market gap: the absence of high-quality Arabic-native Artificial Intelligence systems that capture the linguistic and cultural complexity of the Arabic-speaking world. The team described technical work to solve dialectal variability, including a technique called layer injection that lets the model learn dialectal variations without retraining the entire model. To train and validate Kawn, Misraj said it curated and refined more than 2 trillion Arabic tokens from a wide range of domains and regions, supported by filtering and quality pipelines.

Alongside Kawn, Misraj launched Workforces, an AI platform for creating and managing Artificial Intelligence agents to automate workflows, analyse data and scale operations. The company positioned the platform as especially relevant for teams in operations, customer support, marketing and sales, and said it will pursue sector-specific models tailored to healthcare, legal, insurance, business, education and public services. Misraj also signalled plans for richer multimodal capabilities that combine text, speech and vision. The stated long-term vision is for Kawn to become a primary Arabic Artificial Intelligence foundation across the region, helping ministries, banks, insurers, universities and enterprises deploy solutions that operate naturally in Arabic and integrate with existing systems.

55

Impact Score

How high quality sound shapes virtual communication and trust

As virtual meetings, classes, and content become routine, researchers and audio leaders argue that sound quality is now central to how we judge credibility, intelligence, and trust. Advances in Artificial Intelligence powered audio processing are making clear, unobtrusive sound both more critical and more accessible across work, education, and marketing.

Contact Us

Got questions? Use the form to contact us.

Contact Form

Clicking next sends a verification code to your email. After verifying, you can enter your message.