The world of AI assistants is moving beyond one-off questions and answers. Perplexity is pushing this evolution with Teammates, a new feature that gives AI persistent roles and the autonomy to manage ongoing projects from start to finish.
What Are Perplexity Teammates?
Think of Teammates
as specialized, full-time AI employees rather than a simple chatbot. Announced in early November 2026, this feature allows users to create persistent AI agents assigned to specific, long-term responsibilities. Instead of giving the AI a single prompt for an immediate task, you delegate an entire area of work, such as monitoring a marketing campaign, triaging customer support tickets, or researching competitors. Each Teammate is designed to own its role, proactively working in the background 24/7 without needing constant instruction. It builds on Perplexity's existing 'Computer' platform, which breaks complex requests into smaller steps, but takes the concept much further by introducing autonomy and memory.
The Power of Separate, Monitorable Threads
The core innovation highlighted in the headline is how users interact with and supervise these agents. Each Teammate you create operates within its own dedicated, persistent thread that appears in your Perplexity sidebar. This thread acts as a complete, unbroken log of the agent's activities. You can check in at any time to see what the Teammate is currently working on, what it has completed, and the context it's using to make decisions. This structure provides a crucial layer of transparency and control. Furthermore, each Teammate runs in a completely isolated technical environment. This means every agent has its own virtual desktop, separate browser session, and independent cookies, preventing any overlap or interference between, for example, a marketing agent logged into social media and an engineering agent working in a code repository.
How Multi-Agent Orchestration Works
Teammates don't just work alone; they can manage other AI agents. This concept is known as multi-agent orchestration. When you assign a large, complex goal, a Teammate can analyze the project, break it down into smaller, parallel tasks, and then spin up additional, temporary agents to execute each part. For example, a research Teammate could deploy three other agents to simultaneously investigate different market segments. It then collects the findings from each agent, synthesizes the information, checks the results, and presents a comprehensive report back to you in its own thread. This ability to delegate and coordinate allows Teammates to tackle large-scale projects with a level of efficiency that mimics a human team.
Collaboration and Real-World Use Cases
Perplexity has designed Teammates to integrate into existing workflows. Agents can be added directly into Slack channels, where human colleagues can assign them tasks or ask for status updates. Teammates can also pass work to each other; one agent might find bugs on a website and then hand off its findings to another Teammate tasked with preparing fixes. The potential applications are broad. A sales Teammate could monitor customer interactions and proactively draft follow-up emails. An operations Teammate could track project deadlines and automatically chase missing documents. For developers, an agent could monitor a GitHub repository, respond to review comments, and triage incoming bug reports.
A Shift From Prompting to Delegating
The introduction of Teammates signals a significant shift in how we interact with artificial intelligence in a professional context. The paradigm is moving away from the 'prompt-and-response' model of chatbots toward one of genuine delegation. By giving agents persistent memory, specific roles, proactive monitoring capabilities, and the ability to work around the clock, Perplexity is positioning AI as a true digital workforce. This feature is initially available to users with a Perplexity Max subscription. While the technology is still in its early stages, it offers a compelling glimpse into a future where employees might manage a team of AI specialists, focusing their own efforts on strategy, creativity, and final approval rather than on the minutiae of task execution.
