Perplexity introduces Teammates, always-on AI agents for work

Perplexity is expanding from answer engine to workplace agent platform with the launch of Teammates, a new set of always-on AI agents for teams. According to The Verge, the feature is aimed at the same office-agent territory being pursued by OpenAI and Anthropic, with agents that can hold roles, use existing workspace context, and carry out tasks over time.

For creators, marketers, freelancers and small businesses, the news matters because Perplexity is pitching Teammates less as a one-off chatbot session and more as a persistent digital coworker. The catch: The Verge reports that access requires Perplexity’s $200-per-month Max plan.

What Teammates adds to Perplexity

Perplexity’s Teammates are designed to operate inside a workspace and take on more than one role at once. The company says they can draw on projects, skills and memory from the Perplexity tools a user already has in place.

That is a notable shift from the classic chatbot pattern: ask a question, receive an answer, start again. Teammates appear to be organized around longer-running work, with each agent getting its own thread in the Perplexity app so users can follow what it is doing and continue the interaction in context.

The feature also connects to Slack, according to The Verge, giving teams a way to interact with Teammates where work discussions are already happening. For small teams, that could make the difference between an AI feature that lives in a separate tab and one that becomes part of daily operations.

Perplexity says Teammates integrate with Perplexity Computer, its system for deploying agents on a user’s device to complete tasks. The company describes Teammates as building on that foundation by supporting “ongoing, large-scale work.”

A more agentic workspace, not just search

Perplexity is best known as an AI answer tool, but Teammates pushes it toward a broader productivity role. The key idea is persistence: an agent can be assigned a job, maintain a thread, and continue working rather than simply responding to a single prompt.

That could be useful for recurring research, project tracking, competitive monitoring, content planning or internal coordination — as long as the work fits within Perplexity’s actual capabilities and connected environment. The source does not list specific supported task types, so teams should treat Teammates as a platform direction rather than assume it can replace specialized workflow tools.

The individual-thread model may also help with oversight. If several AI agents are running at once, users need a clear way to see status, review outputs and intervene. Separate threads give each Teammate a visible trail, which is important for teams that want AI assistance without losing track of who — or what — did the work.

How it lines up against OpenAI and Anthropic

The Verge frames Teammates as Perplexity’s response to OpenAI’s Dots and Anthropic’s Claude Duet. We have covered those broader agent moves in OpenAI’s Dots brings coding-agent ideas into ChatGPT and Anthropic launches Claude Duet AI agent in research preview.

The shared direction is clear: major AI companies are trying to turn assistants into always-available workers that can remember context, take action and operate across tasks. For buyers, the practical question is not which agent sounds most ambitious, but which one fits the systems they already use and can be supervised safely.

Perplexity’s version is tied to its own workspace memory, projects and skills, plus Perplexity Computer. That may appeal to existing Perplexity users who already rely on it for research-heavy work. Teams comparing it with OpenAI or Anthropic should focus on workflow fit, admin needs, reliability and the total subscription cost, not just the headline feature.

For readers evaluating Perplexity more broadly, our ChatGPT vs Perplexity comparison may help frame where the tool already differs from a general-purpose chatbot.

What small teams should watch before adopting

The most immediate barrier is pricing. The Verge reports that Teammates requires Perplexity’s $200-per-month Max plan, which makes it a more serious commitment than a casual AI subscription.

Before paying for an always-on agent layer, small businesses and freelancers should ask a few practical questions:

• What recurring work would actually benefit from a persistent agent?
• Who will review each Teammate’s outputs and actions?
• Does the team already use Perplexity enough for its memory and projects to be useful?
• Will Slack access improve collaboration, or create more noise?

Perplexity CEO Aravind Srinivas suggested on X that the company imagines organizations eventually having “hundreds of teammates” for each person. That points to a future where managing AI workers becomes a workflow in itself.

For now, Teammates looks like another sign that the AI assistant market is moving from chat windows toward managed, role-based agents. The opportunity is real for teams with repeatable knowledge work, but so is the need for clear supervision and a realistic view of what these systems are being asked to do.