11:30-12:10

Lightning Pitch

[①Citadel AI]

Trust and Control for Every AI Agent

AI agents act autonomously to plan and take action toward their goals — but that autonomy carries real risk. On top of that, different departments often run different AI platforms, making company-wide governance even harder. In this session, we introduce a new approach to governance: gain visibility and control over every AI agent across the organization — simply connect, and governance begins.

[②Helpfeel]

Common Characteristics of Successful AI Agent Adoptions

The success or failure of adopting AI agents does not depend on the AI itself. What successful cases have in common is the design of work processes and knowledge frameworks that enable the AI to function properly.
In this session, we will introduce key points for enhancing the accuracy and scope of AI applications, leading to measurable impacts such as a reduction in inquiries and increased productivity.

[③Ippu Senkin]

A New Option: On-Premises × Local Generative AI
— A Solution to Token Costs, Geopolitical Risks, and Data Leakage Risks —

Unlike traditional chat-based AI, AI agents autonomously consume large volumes of tokens, leading to a series of cases where API costs in cloud environments have ballooned beyond expectations. If you limit usage out of concern for costs, the benefits of AI are cut in half. With a local AI agent running on-premises, you can make full use of AI without worrying about token costs. In this session, we’ll introduce the cutting edge of the rapidly evolving “on-premises × local AI agent” landscape.

[④ai&]

Leading-Edge AI, Powered by Infrastructure in Japan
— Exploring Sovereign AI Through a Live Comparison

ai& is an AI company focused on inference and data sovereignty, delivering AI from Japan securely and affordably. Today, the main battleground in AI is model development, but as AI agents spread and token consumption explodes, it will inevitably shift in the near future to inference: the continuous use of AI. Through ai& Inference, we offer cutting-edge open models and more at significantly lower cost, in a secure environment where all processing takes place within Japan. Behind this are our own data centers in Japan, GPUs, a heterogeneous architecture that pairs each workload with the optimal hardware, and vertical integration across the entire AI stack. In this session, we will share our inference-first vision for the future of AI from Japan.