Manus: Recent AI Developments Curated by Our CIO
In March, our Chief Information Officer, Sébastien Landeau, highlights significant advancements in artificial intelligence. This month, we focus on two key developments: the debut of a new player in the Large Language Model (LLM) sector and a groundbreaking initiative from OpenAI involving sophisticated intelligent agents.
1/ Manus: A New Entry in the AI Landscape
The landscape of artificial intelligence is experiencing continuous growth, with the latest newcomer being Manus, a Chinese AI. This general-purpose agent is crafted to assist in various tasks, ranging from everyday professional functions to extensive business projects.
Sébastien Landeau, CIO of VirtualExpo Group, emphasizes that Manus is a valuable asset for businesses. He noted that “Manus excels in market analysis, B2B sourcing, and processing commercial data.” This AI streamlines supplier searches, enhances the performance of online stores, and provides robust financial analysis capabilities.
In addition to these features, Manus can optimize service and product comparisons and help structure and manage enterprise data effectively.
Currently, Manus is in a private testing stage. For more details, visit manus.im.
2/ OpenAI’s Revolutionary AI Agents
Meanwhile, OpenAI continues to innovate in the realm of artificial intelligence in the United States. This California-based organization has made new APIs accessible, allowing developers to create intelligent agents capable of executing actions directly on behalf of users.
A New Era in AI Agent Development
The idea of AI agents has been around since late 2024, with many existing initiatives for their development. However, OpenAI is taking a significant step forward by providing an integrated, all-in-one solution.
Before this, developers often struggled to get various tools to work harmoniously. With OpenAI’s new APIs, everything is integrated, offering improved efficiency and compatibility across systems.
Enhanced Autonomy and Proactivity of AI Agents
Sébastien Landeau points out that previously, AI models were limited primarily to responding to queries and offering suggestions. The newly released APIs from OpenAI enable developers to craft proactive AI systems that can make decisions and take actions based on the user’s requirements and their environment. These advancements pave the way for unprecedented possibilities—automating complex tasks, managing tasks across tools, or handling autonomous processes.
Practical Applications
The newly introduced AI agents can perform real-time web searches to gather pertinent information and process knowledge from uploaded documents, applying filters to refine data selection for thorough analysis. Additionally, they can navigate through a computer’s interface to capture screenshots and extract significant information.
Developer-Friendly Tools
To support the development of these intelligent agents, OpenAI has launched two crucial tools. The API Responses tool manages the dynamics between AI and users, while the SDK Agents tool is an open-source kit designed to facilitate agent workflows and the integration of models with various internal systems.
As Sébastien Landeau states, “OpenAI is not merely enhancing its models; it is transforming how we interact with AI. This transition from passive to active AI can change our work habits and interactions with technology.”
For further insights, please visit OpenAI – New Tools for Building Agents.
