Kodamai Launches Mathematically Verified AI Agent Platform, Signs First Mills as First Customer

Kodamai AI agents have entered the enterprise market with a platform the company says can mathematically prove the correctness of every action before execution. The Glasgow and Riyadh-based company emerged from stealth on April 13, 2026, introducing the Kelvingrove platform, which uses Category Theory, Type Theory, and Neuro-Symbolic AI to give each agent a formally verified interface contract.

Alongside the launch, Kodamai said First Mills (Tadawul: 2283), Saudi Arabia’s largest flour milling company, will deploy the platform across its four production facilities. The deal marks the first use of mathematically verified artificial intelligence agent infrastructure at this scale in the Kingdom.

Why First Mills Chose Kodamai AI Agents

First Mills processes thousands of operational decisions daily across supply chain coordination, quality control, and demand forecasting. Many of those decisions have historically depended on manual processes. The Kelvingrove platform will deploy coordinated fleets of AI agents across the company’s operations to manage procurement signals, monitor production outputs, and automate repetitive decision cycles.

Alaa Shousha, Chief Financial Officer of First Mills, said the partnership represents a long-term investment in operational efficiency. “By integrating the world’s first mathematically verified agentic AI, we are not merely adopting new technology; we are establishing a sophisticated governance framework that ensures every operational decision is provably correct and fully auditable,” Shousha said. “This level of precision is critical for optimizing our supply chain and production cycles, directly contributing to our bottom line while reinforcing our role as a cornerstone of Saudi Arabia’s food security.”

The Mathematics Behind the Platform

Kodamai’s Kelvingrove platform rests on three interlocking mathematical disciplines. Category Theory provides a formal structural model of agents, workflows, and system relationships that is independent of any specific implementation. Type Theory assigns each agent a formally typed interface contract, preventing errors from propagating silently through the agent network. Neuro-Symbolic AI integrates large language models into a formally governed process, validating LLM outputs against typed interfaces before they propagate further.

The combination produces what Kodamai calls an Agentic Operating System. The company says it delivers greater agent scalability than competing frameworks, formal governance built into the architecture, and regulatory-grade explainability. Kelvingrove is compatible with any LLM and any enterprise platform, according to the company.

Executive Perspective on Enterprise AI Trust

Maha Achour, PhD, Co-Founder and Chief Executive Officer of Kodamai, said the shift from experimental to operational AI deployments requires a new standard of trust. “We built Kodamai to give enterprises the ability to deploy AI agents they can govern — systems that are not only intelligent but provably correct, explainable, and auditable by design,” Achour said.

The platform’s governance features are designed for industries where errors carry significant financial, operational, or safety consequences. Kodamai specifically cited manufacturing, supply chain, financial services, healthcare, energy, and logistics as target sectors. The business model positions Kelvingrove as infrastructure for organizations moving beyond pilot programs into fully operational autonomous systems.

Alignment with Saudi Vision 2030

Kodamai said the First Mills deployment aligns with رؤية السعودية 2030‘s goal of building a technology-driven economy. Food security is a stated priority under the Vision 2030 framework, and First Mills holds a significant role in the Kingdom’s flour supply chain. The deployment spans all four of the company’s production facilities, making it one of the larger enterprise AI rollouts in Saudi Arabia’s manufacturing sector to date.

Kodamai is headquartered in Glasgow, UK, with team members across the United States, United Kingdom, and Middle East. The company was founded by three mathematicians and scientists whose research underpins the platform’s formal verification approach. Further details are available on the company’s website.