Openai has launched the chatgpt agenta Agricultural tool capable of autonomously performing complex tasks dividing them into several steps without intervention of the person who orders it. Among other things, the chatgpt agent can navigate on web pages, execute code, analyze data, search and display documentation and generate presentations.
Its function is to perform tasks that an auxiliary or assistant worker would do, but in the form of a digital partner who works with a virtual computer, apart from the work he does who uses it with his computer, and then communicate his results.
This agent is autonomous in the management of workflows handling different tools, but without losing the context of everything he does while performing the operation that has been asked. To do this, use the tools already available by chatgpt. As Operator, which allows the agent to interact with the websites as if it were a human user.
In addition, it has Deep Research, another function that allows you to collect information and write detailed reports. Also Connectors, a utility that offers what is necessary to access different online services, such as Gmail, Google Drive or Dropbox.
The chatgpt agent uses all these tools in a single environment, which includes a finish, web browsers and a file system. The functions of these tools allow the agent to perform tasks with complex workflows and composed of several phases.
Thus, for example, you can download files, make changes to them through a terminal and review the results obtained, all without losing context. Search and use a Google Drive document, access its contents, generate a report with pre -established format from them, compare it with other documents, create quotes in the calendar and add it to these appointments. Also, during meetings, summarize what is in question in them and then save them in a file at the point where you want.
As for online purchases, you can search for the desired products, Buy them by comparing prices and carry out all operations associated with the purchase. According to Openai, the time that the chatgpt agent may take to perform the tasks that are proposed may vary, although they will generally be ready in a few minutes. However, the most complex can take between 15 and 30 minutes to be ready.
For now, Openai Agent will begin his deployment gradually. At first it will be available for Chatgpt Pro subscribers, with 400 questions per month. It will be followed by users of the Plus and Team plans, with 40 monthly Queries. After their arrival at these plans they will have those who have a subscription of the Enterprise and Education plans. For now it is unknown if it will reach users of the free version of Chatgpt, and if so, when it will come and under what conditions.
To access the chatgpt agent, enough Open the tools drop -down menu in chatgpt and choose the agent mode option. With it, users will have access to a screen narrative, as well as the possibility of taking manual control of what the agent is doing at a given time, in case they are making mistakes or is not satisfied with the partial results achieved by him.