Perplexity recently announced the launch of a suite of personal AI agents called "teammates," designed to autonomously handle ongoing tasks with minimal supervision. They operate on the company's cloud assistant platform, Perplexity Computer, which can complete multi-step tasks for users. Perplexity stated that these new agents are built on Computer and bring "additional capabilities for scalable, long-term work."
Users can assign their own roles or tasks to each teammate, allowing them to execute simultaneously while coordinating and managing other teammates to ensure work progress. These agents can also introduce new companions and adjust their tasks based on the workload assigned. Teammates will operate continuously for 24 hours and will notify users when decisions are needed. Each teammate will have a dedicated entry in the sidebar of the Computer interface, making it easy to track their progress and completed work.
This feature is currently available only to Perplexity Max subscribers on the web and iOS platforms. This positioning makes the overall solution relatively expensive, with a monthly fee of US$200 and an annual fee of US$2,000. Users can request agents to alternate tasks, switch focus, or check their status updates during ongoing tasks. This strategy emphasizes proactive collaboration and long-term background operation compared to personal agents from other companies, meaning users can receive continuous work support.
Each teammate will display its current tasks, completed milestones, and next steps in the sidebar of Computer. This design helps users quickly identify moments requiring decisions and maintain transparency in a multitasking environment. For technical workers, content creators, and business decision-makers, this type of automated background collaboration has the potential to enhance efficiency, but it also raises considerations regarding data and privacy.
Evaluation of the Positioning, Availability, and Cost of Next-Generation Personal Agents
In terms of background information, other market trends also indicate that various AI providers are launching agents capable of automating task management and enabling cross-platform collaboration through cloud mechanisms. For example, Anthropic and Claude Duet promote "proactive" agents, emphasizing daily task lists and automatically suggesting action items based on connected applications. This design reflects the trend that "agents not only execute commands but also proactively initiate collaboration and communication." If you consider long-term complex workflows, these agents can significantly reduce the burden of human decision-making and provide real-time status feedback.
Using Perplexity's official positioning as an example, teammates are realized through multi-agent collaboration, allowing them to divide tasks, monitor each other, and introduce new members to expand work capabilities. Agents will operate 24 hours a day and will notify users when decisions are needed. This structure is expected to enhance efficiency in scenarios such as data analysis, content creation, software development, and project management, but it also raises issues regarding data sources, cross-application data flow, and access control. If you are a business decision-maker, you need to assess the compatibility of existing workflows, data security risks, and long-term cost benefits before deployment.
Additionally, the pricing model itself is worth noting. The Perplexity Max plan has a threshold of US$200/month or US$2,000/year, which affects whether teams of different sizes can easily access it, depending on their workload, the complexity of tasks to be automated, and their reliance on stable background collaboration. In comparing pricing with similar competitors, users must weigh the actual value of the agents against their acceptance of privacy and data sharing to determine whether to invest long-term.
For broader promotional phases, businesses and developer communities may focus on the scalability of such agent systems, cross-platform integration capabilities, and the degree of intrusion into existing workflows. Of course, aside from cost factors, the user experience of member management, task assignment, and monitoring interfaces for the agents is also key to assessing real benefits. If you want to test this technology in your workflow, it is advisable to start with a small number of tasks as a pilot and set clear authorization, review, and exit mechanisms to avoid data flow exceeding your control.

