Perplexity launches AI teammates for ongoing work inside Computer

Perplexity has added autonomous AI teammates to its Computer assistant, available only on the $200 monthly Max plan. The move shifts Perplexity from answers to persistent, parallel task management.

What happened

On November 6, 2026, Perplexity launched personal AI agents called "teammates," designed to tackle ongoing work autonomously with minimal supervision. The agents work inside Perplexity Computer, the company's cloud-based agentic assistant that can perform multi-step tasks. Perplexity says the new agents build on Computer and come with "additional capabilities for ongoing, large-scale work."

Users can assign each teammate its own role or task, and teammates can perform those tasks simultaneously. They work together, can manage other teammates to make sure tasks are getting done, and can bring in more teammates and plan the newcomers' tasks accordingly as more work is assigned.

The feature is exclusively available to Perplexity Max subscribers on the web and on iOS. The Max plan costs $200 a month or $2,000 a year.

Key facts

• Teammates are personal AI agents built on Perplexity Computer and designed for ongoing work autonomously with minimal supervision.
• Each teammate can be assigned its own role or task and can perform tasks simultaneously with others.
• Teammates can manage other teammates and bring in more teammates, planning their tasks as workload grows.
• The agents work 24/7 in the background and notify users when decisions are needed.
• Access is limited to Perplexity Max subscribers on web and iOS; the plan costs $200 a month or $2,000 a year.

Our analysis

The launch moves Perplexity further from its search-answer roots into sustained, multi-agent work orchestration. Instead of a single prompt producing a one-off output, teammates create persistent sidebar entries in Computer where an operator can see what each agent is doing and has finished, hand off more work, tell them to switch gears, or request status updates. For teams that already run repetitive content, research, or administrative workflows, this suggests a shift from supervising individual AI tasks to managing a small always-on agent workforce.

The $200 monthly price point is significant. It likely filters out casual users and positions the feature for professionals who can justify the cost with labor savings or expanded output. At the same time, because teammates can manage other teammates and recruit additional teammates, the system is designed to scale with workload rather than only answer discrete questions. That creates a different operating need: operators must be prepared to review autonomous decisions and set clear boundaries, since the agents run in the background and surface decisions only when they need input.

For social media managers, creators, and community operators, the value depends on whether recurring tasks can be broken into roles and handed off with minimal supervision. The ability to run teammates simultaneously and have them coordinate suggests potential for parallel content research, drafting, monitoring, and task tracking, but the source does not specify Perplexity's accuracy, platform integrations, or guardrails. Those gaps should be tested before committing to the Max plan.

What it means for operators

• Map your recurring workflows before paying. Identify tasks that can be assigned as roles, run 24/7, and require minimal supervision. If you cannot define clear handoffs, the $200 monthly cost may not be justified.
• Use teammate threads as a task dashboard. Check each teammate's sidebar entry to see current and finished work, and use status update requests instead of micromanaging single prompts.
• Plan for decision notifications. Since teammates work in the background and notify you when decisions are needed, set aside time to review those prompts quickly so autonomous work does not stall.
• Verify platform access before relying on teammates. The feature is on web and iOS at launch, so operators whose workflows require other surfaces should confirm whether that fits their existing setup.