Fujitsu has introduced its Fujitsu Kozuchi AI Agent, an AI service capable of autonomous operation and collaboration with humans on high-level tasks.
The Fujitsu Kozuchi AI Agent aims to transform business operations by autonomously participating in meetings, analysing data, and proposing solutions. This service is part of Fujitsu's Data Intelligence Platform as a Service (PaaS) and the wider Fujitsu Uvance business model, which intends to deliver cross-industry solutions to societal issues.
Initially, the AI agent will focus on business profitability discussions and negotiations, with plans to extend its application to production management and legal affairs during fiscal year 2024.
Mahesh Krishnan, Chief Technology Officer Oceania at Fujitsu, highlighted the importance of solving key challenges related to AI development. "The future of AI hinges on solving key challenges like resource scarcity and maximising human-machine collaboration. Fujitsu is addressing both with our new AI computing broker middleware and Fujitsu Kozuchi AI Agents. The middleware provides a powerful solution, allowing businesses to better leverage their existing GPU resources and break through processing constraints," he said.
The integration of AI services into Fujitsu's broader digital offerings, such as Fujitsu Uvance's Work Life Shift, is also planned. The use of such generative AI models requires significant investment and expertise, and Fujitsu Kozuchi AI Agent is designed to mitigate these challenges by autonomously thinking and suggesting solutions.
The AI agent employs Fujitsu's proprietary processing logic to dissect abstract issues from discussions into actionable tasks, then executes these tasks using multiple optimal AI models. These models include Fujitsu-developed models like Takane, a language model with notable Japanese language capabilities, and various models from other developers.
An example of the AI agent's application is its role in business meetings, where it can independently participate and analyse financial data. The agent, upon identifying that sales revenue in a specific region has decreased, will autonomously generate and present a data analysis, such as a bar graph comparing sales revenue across different periods. Such functionality supports informed decision-making and streamlines meeting productivity.
Fujitsu's initiative reflects its commitment to enhancing business efficiency and facilitating business transformation through the use of generative AI models, the company shared in a statement.
