AI companies want you to stop chatting with bots and start managing them

On Thursday, Anthropic and OpenAI introduced groundbreaking products that signal a shift in how we interact with artificial intelligence. Rather than engaging in conversations with a solitary AI assistant, users are now encouraged to oversee teams of AI agents that collaboratively handle tasks in parallel. This transition, highlighted by the simultaneous launch of new tools, reflects a broader industry trend away from conversational AI towards a model where AI functions as a delegated workforce. This evolution comes at a time when the concept has reportedly contributed to a staggering $285 billion decline in software stocks, raising questions about the practical effectiveness of this supervisory approach. Currently, AI agents still demand significant human oversight to identify and correct errors, and independent assessments have yet to establish that these multi-agent systems consistently outstrip the performance of a single developer. Despite these challenges, both companies are fully committed to the potential of AI agents. Anthropic's latest offering, Claude Opus 4.6, enhances its most advanced AI model with a feature known as “agent teams.” This allows developers to deploy multiple AI agents that can autonomously divide tasks, collaborate, and operate simultaneously. The interface resembles a split-screen terminal environment, enabling developers to easily navigate between sub-agents and oversee their activities, particularly for tasks involving extensive reading, such as code reviews. This feature is currently available as a research preview. Meanwhile, OpenAI unveiled Frontier, an enterprise platform designed to “hire AI co-workers” for various tasks traditionally performed on computers. Each AI agent in Frontier is assigned a unique identity, along with designated permissions and memory, allowing seamless integration with existing business systems, including CRMs and data management tools. "Our core mission is to evolve these agents into genuine AI collaborators," stated Barret Zoph, OpenAI’s general manager of business-to-business solutions, in an interview with CNBC.