OpenAI has launched ChatGPT Agent, a tool designed to complete various complex tasks autonomously. However, early users indicate that while it shows promise, it is marred by significant performance issues and limitations, reports 24brussels.
Launched on July 18, 2025, ChatGPT Agent merges elements from the earlier tools Operator and Deep Research. Despite the high demand for this service, users have reported delays in its rollout due to unforeseen popularity. Initial feedback from users highlights the system’s sluggishness, glitches, and occasional reliability issues when handling tasks.
The agent requires users to activate it by entering “Agent Mode,” which prompts a list of suggested tasks. Among these are comparisons of products or preparation of personalized lists. However, reports suggest that the agent is slow, often taking an excessive amount of time—like 50 minutes—for tasks that a human could complete in much less time.
In a demo session with OpenAI representatives, product lead Yash Kumar emphasized that the focus is on optimizing for complex tasks rather than speed, advocating that the tool allows for multitasking as users engage in their primary activities.
Kumar commented, “Even if it takes 15 minutes or half an hour, it’s quite a big speed-up compared to how long it would take you to do it.” He also noted the new “Watch Mode” feature, which requires users to monitor the agent while it operates on potentially sensitive tasks, addressing security concerns.Upon initial interaction, users demonstrated how the ChatGPT Agent could assist with finding products online, such as searching for and comparing flower delivery options.
Yet, the agent’s inability to execute transactions directly has also raised concerns. When asked to make a purchase, the AI claimed it could only guide the user through the ordering process without direct engagement, a limitation that frustrated potential users relying on its promised functionality. This inability to perform key tasks raises questions about the viability of ChatGPT Agent as a reliable assistant for everyday tasks.
Despite some operational deficiencies, the ChatGPT Agent still provides valuable information to assist users in making decisions, like comparing products or generating persuasive messages suitable for communication. However, its current lack of capability to perform transactions or engage with third-party websites limits its appeal.
In closing, the ChatGPT Agent represents a notable advancement in AI-assisted applications, albeit one accompanied by significant limitations that OpenAI must address for the tool to fully meet user expectations. The road ahead will require continued optimization and development to enhance functionality and user experience.
ChatGPT Agent could pave the way for more sophisticated virtual assistants, provided it can resolve its current issues and fulfill its experimental potential.
