Perplexity Launches Teammates Multi-Agent Platform for Enterprise Project Management

Teammates is built on Perplexity's previously launched "Computer" system. The latter can invoke AI agents to perform specific tasks on a user's device, while Teammates further extends this capability, no longer limited to single-task processing but focused on managing and advancing long-term, large-scale work projects. The system can coordinate multiple AI agents working together, allowing different agents to simultaneously handle the work content each is responsible for.

Similar to OpenAI's Dots and Anthropic's Claude Duet, Teammates adopts a multi-agent collaboration model. Users can assign different responsibilities to each agent, such as marketing, sales, customer support, operations management, or financial analysis. Multiple agents can execute tasks in parallel and collaborate with one another based on overall project progress.

Perplexity says Teammates is not merely a tool that automatically executes commands, but can continuously take responsibility for advancing projects. The system plans work content, tracks execution status, and proactively takes action based on changes in the environment. For example, in a sales scenario, an agent can discover that a long-term customer has stopped placing orders, then proactively analyze order history and after-sales records, investigate the cause and propose solutions, and even generate discount offers or remind the team to intervene.

In terms of product design, each Teammate has an independent and persistent conversation thread. Users can view its work status, completed tasks, and latest progress at any time, and adjust the direction of tasks at any moment. The system also has memory capabilities, able to record project goals, execution history, and shared file contents, so that subsequent tasks build on previous work results without the need to frequently re-explain background information.

Perplexity has also incorporated Slack integration into the product system. Enterprise employees can interact directly with AI agents in Slack channels, assign them tasks, check progress, or obtain results. At the same time, different agents can also hand off work to one another. For example, after one agent discovers a vulnerability in a website, it can transfer the investigation results to another agent responsible for developing a remediation plan.

Perplexity CEO Aravind Srinivas said the company's long-term vision is for every employee to have a large number of AI colleagues around them. He believes that in the future, everyone in every organization may have hundreds of dedicated agents assisting with daily work, thereby freeing human employees from a large amount of repetitive coordination work.

However, compared with competitors, the product Perplexity launched this time is not inexpensive. Teammates is currently available only to Perplexity Max subscribers, and that subscription plan costs $200 per month. This makes it one of the higher-positioned enterprise-grade AI agent products on the market.

From a strategic perspective, the launch of Teammates marks Perplexity's acceleration of its transformation from a company known for AI search and question-answering services into a platform provider targeting the enterprise productivity and work automation market. As OpenAI, Anthropic, and other AI companies roll out agent systems capable of continuous operation, competition around "AI colleagues" and "digital employees" is rapidly heating up.

It is widely believed in the industry that the focus of competition for a new generation of AI products is shifting from chatbots to agent platforms that can autonomously execute tasks, manage projects over the long term, and deeply integrate with enterprise software. Perplexity's launch of Teammates is an important move under this trend, with the goal of gaining more share in the enterprise AI automation market.