OpenAI, launches "Workspace Agent"…… ChatGPT expands into enterprise business automation

OpenAI has expanded the application scope of “ChatGPT” beyond the existing conversation interface. The newly announced “WorkSpace Agent” is a feature designed to help enterprise users automate repetitive tasks, with the core capability to continue running in the cloud even when employees are offline.

OpenAI publicly introduced WorkSpace Agent on the 24th via a blog post. The feature is based on the Codex model and is not simply a question-and-answer chatbot, but operates more like a “continuous human assistance.” It states that it can connect to external software, maintain context across projects, and execute multi-step tasks without needing repeated instructions.

The newly launched feature is closer to an extended version of the existing “GPTs.” GPTs are customizable ChatGPT functions that users can tailor to specific tasks, interests, or workflows without additional coding. WorkSpace Agent takes this further, focusing on automation aligned with real organizational workflows. It can handle tasks such as report writing, code generation, message responses, and supports internal sharing within the organization for team collaboration and improvement.

The usage is relatively straightforward. Users simply describe the desired workflow in a new tab within the ChatGPT interface. ChatGPT then designs the necessary programs, connects relevant tools, and tests until completion. If the user is satisfied with the results, they can activate the Agent to execute tasks on schedule or respond automatically under certain conditions.

The enterprise-level “Agent AI” competition officially begins

OpenAI’s release comes at a time when the enterprise “Agent AI” competition is intensifying. Against the backdrop of major tech giants like Google, Microsoft($MSFT), and Amazon Web Services continuously investing heavily in building autonomous business systems, OpenAI feels a strong push from Anthropic.

Market commentary notes that recently, Anthropic has rapidly increased its presence in the enterprise AI field with tools like Claude Code and Co-work. Holger Mueller of Constellation Research believes that OpenAI was once “the center of AI innovation,” but now has fallen into a position of rapid follower. He states that Anthropic is shifting the focus of AI competition from consumers to enterprises and has gained dominance.

This assessment explains why OpenAI is pushing WorkSpace Agent to the forefront. Generative AI has surpassed the stage of merely enhancing individual productivity; industry demands its evolution to connect with actual enterprise approval processes, team collaboration, and work handovers. OpenAI recognizes this and states that the design goal of WorkSpace Agent is to gather context from multiple systems, follow team workflows, and request approval when necessary.

Security concerns still persist… Equipped with approval processes and access restrictions

However, the enterprise sector remains cautious about adopting autonomous AI. Security issues such as prompt injection attacks or data access permission abuses are seen as major obstacles. OpenAI is aware of these concerns and has announced that enterprises can restrict the data access scope and available tools for the Agent, and provide features to require pre-approval for sensitive tasks.

This is not just a convenience feature but a necessary condition to gain trust in the enterprise market. Since automation is connected to operational systems, the performance of the Agent is important, but controllability and audit traceability are also becoming key competitive factors.

Currently, WorkSpace Agent is available as a “research preview” to ChatGPT Business, Edu, Enterprise, and Teachers subscription users. It is free until May 6, after which it will switch to a points-based billing system. OpenAI states that it will continue supporting existing customized GPTs and plans to offer a future feature to convert them into WorkSpace Agents.

This release is expected to serve as a watershed in assessing whether OpenAI can surpass consumer-grade AI giants and establish a foothold as a provider of enterprise productivity platforms. Ultimately, market judgment will depend on whether the AI can be safely and stably deployed in actual business operations, rather than how “smart” it is. Therefore, initial reactions to WorkSpace Agent are highly anticipated.

TP AI Notes This article uses a language model based on TokenPost.ai for summarization. The main content of the text may be omitted or not fully aligned with facts.