Perplexity Launches Always-On AI Agent 'Teammates' at $200 Monthly

What happened

Perplexity has launched "Teammates," a set of always-on AI agents that carry on work across ongoing projects, each running on its own virtual machine. The Decoder reported the launch under the headline "Perplexity now has its own always-on AI agent."

Access depends on the Max plan, which costs $200 per month. According to the same reporting, that price sits well above comparable offerings from OpenAI and Anthropic, so the feature is positioned at the premium end of the market rather than as a general-purpose add-on.

The launch arrives in a crowded field. Always-on, persistent agents are becoming a product category in their own right, and the move signals that Perplexity intends to compete on continuous, project-based work rather than only on answering individual queries.

Why it matters

The significant shift here is from AI as a tool that responds on demand to AI as a participant that keeps working between sessions. An agent that holds a project open, runs on dedicated compute and picks up where it left off changes what organisations can delegate. Research, monitoring, document preparation and multi-step workflows that previously needed a person to restart them each time can, in principle, run continuously. That affects how teams are structured, how work is handed over, and where human review is placed.

For leaders in experience and digital transformation, the question is less whether the technology works and more where it earns its place. A premium price point raises the bar for demonstrable return. Organisations will need clear criteria for which tasks an always-on agent should own, how its outputs are checked before they reach customers, and who is accountable when it acts on a project overnight.

The Renascence take

Most coverage will focus on the feature list and the price. The more useful lens is the handover. Every always-on agent quietly moves work from a person who starts a task to a system that continues it, and that is a service-design decision as much as a technology one.

The features will matter less than the handoff points. Customers and employees experience an agent through the moments where it stops, asks, or reports back, so those moments deserve the same design care as any frontline journey. Operators should map each delegated workflow for trust, not just throughput, and decide in advance what a good interruption looks like. Paying a premium for persistence only pays off if the people relying on the output can see what happened and why.