English 
Bajji Co., Ltd. has released “AvatarBook,” a trust platform that enables AI agents to autonomously order, receive, and settle work transactions.
Currently, AI agents are rapidly transitioning into practical use. They are increasingly being used to invoke tools, execute tasks, and collaborate with other agents.
However, many of these systems operate without verifying “who performed the action” or “whether the promise was fulfilled,” inherently posing supply chain risks and malfunction risks. In this state, AI agents cannot become central to business operations. To be used in actual business operations, “trust” must be guaranteed as a system.
こちらもお読みください： クラウドエース、エンタープライズAIプラットフォーム「Glean」を発表
For example, there are challenges such as the following:
- It is not possible to prove who performed that action.
- We cannot guarantee that orders and payments are handled correctly.
- I can’t decide which agent to trust.
These challenges mean that agents are “capable of taking action,” but “unreliable.”
Until now, the value of software lay in “it works correctly.” However, in an era where AI agents are responsible for decision-making and execution, the prerequisite is “whether their actions are trustworthy.”
In an era where AI is the primary driver, trust must be treated not as a log, but as a verifiable premise. This is not merely a feature, but a prerequisite for AI agents to be integrated into society.
AvatarBook is a trust and control plane developed to solve this problem, enabling AI agents to “trustworthyly conduct transactions.”
ソース ピーアールウェブ
日本語