Fujitsu’s Multi-AI Agents to Boost Supply Chain Resilience

Fujitsu has announced the development of a new multi-AI agent technology designed to optimise supply chain efficiency. By securely connecting AI agents across multiple companies, the system helps businesses react quickly to fluctuating market conditions and operational changes.

The technology will undergo joint trials to test its security and agility, with Fujitsu planning future expansion into new industries.

As a multinational information and communications technology company, Fujitsu provides a range of digital transformation services and solutions. Its work centres on five key technologies: AI computing networks, data and security and converging technologies.

This new collaborative tool allows different companies within a supply chain to respond promptly to changing circumstances by considering the status of other AI agents in the network.

This process could help businesses maintain agility and protect their supply chains from disruption. Early tracking of potential issues allows business leaders to make informed decisions to mitigate risk.

Testing AI for supply chain optimisation

Beginning in January, Fujitsu will initiate field trials of the technology to help increase supply chain optimisation at Rohto Pharmaceutical Co., in collaboration with a leading Japanese university. The multi-agent technology is designed to streamline regular daily supply chain operations. A key asset is its potential ability to facilitate fast recovery during sudden demand shifts or external disasters that impact the business.

After the initial trials with Rohto Pharmaceutical, Fujitsu intends to continue its technological enhancements, aiming for expansion into sectors such as manufacturing. Fujitsu will also work to implement the technology for more complex supply chains.

This could assist with global supply chain transformation. The technology has the potential to enhance supply chain resilience and support sustainable business operations.

In line with its Uvance® business model, Fujitsu's technology can implement secure data collaboration, with AI agent cooperation across industries and borders. This could lead to more resilience, greater reliability and improved governance in multi-vendor environments.

Advancing industry with agentic AI

“A leading Japanese university is actively promoting Cyber-Physical Systems (CPS) research and working to improve efficiency across the industrial value chain," explains Katsuki Fujisawa, Professor, Digital Twin Research Unit at the university.

"Moving forward, by collaborating with Fujitsu's agentic AI technology to optimise the supply chain, we aim to contribute to the advancement of industry and the resolution of societal challenges.”

Fujitsu's new system features two main components: a global optimal control for AI agents and a secure inter-agent gateway.

The optimal control technology approximates the preferred conditions for a partner company using limited information, allowing it to identify the overall optimal state for the supply chain quickly.

This enables effective collaboration between various agents from different companies without requiring the exchange of confidential data.

Secure and seamless AI collaboration

The Fujitsu secure inter-agent gateway is designed for seamless and secure collaboration among external AI agents, protecting confidential information while enabling distributed AI training.

During the setup phase, AI agents learn supply chain characteristics by leveraging knowledge distillation, where data is drawn from multiple 'teacher models' into a single 'student model'.

While in use, this model is able to uncover malicious attacks and prevent the spread of confidential information. It achieves this by repeatedly simulating AI agent behaviour and updating information in a safe and secure format.