Humain invests in Tarjama and Arabic.AI as linguistic-sovereignty bet shifts from training general models to engineering dialects and corporate documents
Humain, backed by the Saudi Public Investment Fund, announced a undisclosed strategic investment in the Arabic.AI platform and its parent company Tarjama, based in Dubai. The alliance focuses on developing an integrated system for Arabic translation management, digitising archives and old documents, and building intelligent agents to examine and draft contracts, tenders and regulatory files within the Saudi market and beyond.
The investment reveals a fundamental shift in the region’s AI infrastructure strategy; the Saudi company, whose board is chaired by the Crown Prince, preferred acquiring about eighteen years of specialised expertise rather than trying to build it from scratch. Tarjama, founded by Noor Al-Hasan in the UAE in 2008, has a record of processing more than two billion words through its translation memory, covering over 100 language pairs, with native support for roughly 22 Arabic dialects, including Modern Standard Arabic, Gulf, Egyptian, Levantine and Maghrebi, and serving more than 700 institutional and government clients across thirty countries.
The existing gap between general language models and audit-ready enterprise systems is the primary driver behind the deal.Humain already has its flagship model ALLaM, trained on more than 500 billion Arabic tokens, but smooth conversation in Modern Standard Arabic does not necessarily mean the model can handle handwritten text in an old property register, understand a legal clause phrased in a mix of local administrative dialect and commercial terminology, or meet compliance requirements of banks and regulators that require a human reviewer in the audit workflow.
The two companies plan to launch an initial suite of three core products: an Arabic-specific translation-management system built on the Cleverso engine, a tool to digitise paper archives and convert handwritten or damaged texts into searchable Arabic data, and AI agents to read, review and audit government contracts and tenders. The joint solutions are slated for official unveiling at the LIP 2026 conference in Riyadh.
This trajectory reshapes the technical project priorities for companies and organisations operating in the Gulf, Egypt and the region at large; it shows executives and data teams that relying solely on importing or training massive models with high compute costs does not solve complex operational challenges. Real value in the regional enterprise environment now demands tools that understand dialectal diversity and everyday transaction contexts, compelling digital-transformation teams to build robust document-audit pathways rather than relying on ready-made conversational interfaces.