OpenAI debuts ChatGPT agent, pushing the tool beyond its traditional ‘chatbot’ role to now execute complex tasks for users. As the company describes it, “ChatGPT now thinks and acts, proactively choosing from a toolbox of agentic skills to complete tasks for you using its own computer.”
Offering a glimpse into its new capabilities, OpenAI says the agent can now manage tasks such as planning and buying ingredients for a family breakfast, reviewing a user’s calendar to brief them on upcoming client meetings, and analyzing competitor companies to prepare a slide deck.
As of this Thursday, these agentic capabilities are available for Pro, Plus, and Team users. OpenAI calls this launch “just the beginning,” promising regular improvements.
ChatGPT agent can be activated directly via the tools dropdown in the composer when users select ‘agent mode’ in any conversation at any point in time.
A unified agentic system powers these advanced features, making them useful in both personal and professional contexts.
But with new perks come new risks. OpenAI acknowledges that, despite built-in safeguard measures for handling sensitive user data, there is still an increased amount of risk associated due to the agent’s capability to have direct access to a user’s data through connectors or websites where the user has logged it in.
Currently in beta, the ChatGPT agent may still produce results that feel “rudimentary in formatting and polish,” OpenAI admits. There are also occasional lags between the viewer and exported PowerPoint slides—an issue the company is actively working to resolve. Development is already underway to train the next version of ChatGPT’s presentation tool to deliver more refined and professional outputs.
TechIntelPro empowers B2B technology decision-makers with the latest industry trends, deep insights, leadership perspectives and actionable intelligence to drive both organizational and professional milestones—and thrive on the tech pulse.