In Benchmark Tests, MetisJean Reduced Token Usage by 90%
Dubai, 1 October, 2026 — MetisJean, an ambitious tech startup, has launched in the United Arab Emirates with the goal of supplying the nation’s public and private sector organisations with sovereign AI capabilities. The firm is focused on generating 100% of its revenue and growth in the UAE, has ambitions for rapid scaling in the region, and aims to play an active role in the region’s adoption of artificial intelligence.
MetisJean is the research and development company behind Jean, a sovereign AI operating system that lowers the entry barrier to genuine AI ownership. Rather than renting intelligence from external models hosted abroad, organisations run Jean inside their own infrastructure, where she connects to all internal business systems (applications, messaging, databases, etc) and learns how the business works.
Jean delivers significant cost savings and reductions in carbon emissions: in benchmark tests, Jean reduced token usage at LLMs by 90%*. Each time a business presents a new request to an LLM through Jean, she replicates the code and the workflow and uses these for future repeat requests. This process eradicates the need for multiple requests to external LLMs.
Because Jean draws directly from the organisation’s data and intelligence, she eliminates the risk of hallucinations. If Jean does not have the correct information or is unable to provide it because of restrictions related to an employee’s role, she will simply inform the user.
Alex Haywood, CEO and Founder of MetisJean, said: “The Large Language Models (LLM) are so complex, with billions of inputs and interacting parameters, that even experts are unable to say how an LLM has arrived at a particular response. This raises questions about decisions and accountability – particular with the emergence of autonomous AI agents. In contrast, Jean aims to provide a ‘white box’ for decision making. Businesses and their employees can verify how they have reached every conclusion.”
Jean governs every action in line with company policies, with industrial, national and global regulations, and in both English and Arabic.
Basing MetisJean in the UAE is a deliberate choice due to the strong alignment between the National Strategy for Artificial Intelligence 2031 and the team’s motivations for building Jean.
“The UAE’s AI strategy is far sighted in guarding against some of the risks posed by the technology – risks such as people and businesses training their AI replacements; diminishing professional capabilities through ongoing AI usage; and an over-reliance on external actors. All it takes to derail a business is an AI vendor increasing their fees, changing the functionality or removing a feature,” Haywood explained.
The team designed MetisJean to enable businesses to harness the capabilities of artificial intelligence without compromising privacy or leaking proprietary data.
All of the company’s growth in the coming years will come from the UAE, either from organisations based here, or foreign firms doing business here and needing assistance with sovereign AI.
Haywood believes that MetisJean is particularly suited to the demands of UAE organisations for sovereign AI capabilities. He commented: “We are prioritizing UAE industries with strict data residency and compliance requirements – such as finance, healthcare, energy, and government. Black-box, foreign-hosted LLMs are a liability for businesses in these sectors.”
MetisJean raised $3 million (AED 11m) in pre-seed funding to build a functional product that is now shipping to its first customers. With ambitious plans for further build out of the product, the startup is now seeking $5m (AED 18.4m) of further investment from funds in the region.
— ENDS —
*= Benchmark tests occurred in a controlled environment and results will be available for peer review in a forthcoming whitepaper. We can provide details to journalists upon request.
About MetisJean
Headquartered in Dubai, MetisJean is the research and development company behind Jean, a sovereign AI operating system that lowers the entry barrier to genuine AI ownership. Rather than renting intelligence from external models hosted abroad, organisations run Jean inside their own infrastructure, where she connects to all internal business systems (applications, messaging, databases, etc) and learns how the business works.
By building unique intelligence within the confines of each customer, MetisJean guards against the risks posed by externally hosted LLMs, such as training the models to replace businesses and roles, reduced professional competence through regular AI usage and over-reliance on external actors.
MetisJean has ambitious plans to scale in the UAE and play an active role in the region’s adoption of artificial intelligence.