✉ The Friday AI Brief: the week's 5 best AI stories, tools & comparisons — in your inbox every Friday morning.

Perplexity AI logo

Perplexity Launches Teammates: Persistent AI Agents That Take On Ongoing Work — for $200 a Month

Perplexity has launched Teammates, persistent AI agents that take a high-level goal and keep working at it — with your projects, memory and files already loaded — instead of starting fresh from a prompt each session. Why it matters: Perplexity is now selling the same always-on coworker idea as OpenAI’s Dots and Anthropic’s Claude agents, and it is charging $200 a month for it.

November 5, 2026

What happened

Perplexity launched Teammates on November 5, 2026, announced by chief executive Aravind Srinivas. The feature builds on Perplexity Computer, the multi-model agent system the company launched in February 2026 for its Max tier. The difference is persistence: Computer worked in temporary sessions, while a Teammate is a named agent with an assigned role that sticks around and owns ongoing work end to end.

Each Teammate inherits what already exists in your Perplexity workspace — your Projects, Memory, Skills, connectors, files and past sessions — so it does not need to be re-briefed every morning. Perplexity’s examples include triaging bugs, running QA loops, monitoring security, analyzing data and managing A/B tests across marketing, sales, support and operations work, with suggested roles such as an on-call engineer, a legal research assistant and a storefront manager.

How Teammates work together

You can run several Teammates at once on different jobs, and they can pass tasks to one another and exchange messages through Slack. Adding a Teammate to a Slack channel lets coworkers hand it work directly, while it watches the channel and acts when something changes. Each Teammate runs in its own isolated desktop and browser environment, and you can review its work, take over its desktop, or redirect it. Perplexity says the agents learn from corrections: tell one to handle a recurring situation a certain way every time, and it keeps doing that without being asked again.

Srinivas said Perplexity’s own internal workspace has used about 10,000 teammates within the last 30 days, according to The Verge’s reporting.

Price and availability

Teammates is rolling out first to Max subscribers on the web, in what Perplexity describes as a research preview. Max costs $200 per month; free and Pro users have to upgrade. Native mobile and desktop apps are coming, according to Srinivas. That price puts Teammates well above mainstream chatbot plans and squarely in the premium agent tier alongside OpenAI’s and Anthropic’s work agents, though those rivals mostly target organizations rather than individual subscribers.

Why it matters

The agent race is shifting from “ask a question, get an answer” to “hand over a responsibility.” Teammates is Perplexity’s version of that shift, and its bet is that context is the product: an agent that already knows your projects, files and past sessions is worth more than a smarter model starting from zero. The open question is the same one facing every always-on agent launch this year — whether subscribers will pay $200 a month for background labor they still have to supervise. Perplexity’s internal usage figure suggests heavy use inside at least one company; whether that transfers to paying customers is what the preview will test.

Frequently asked questions

What is Perplexity Teammates?

A set of persistent, always-on AI agents built on Perplexity Computer. Each Teammate takes a role and a high-level goal, then works continuously using your existing projects, memory, skills and files.

How much does it cost?

Teammates requires the Perplexity Max plan at $200 per month. It is rolling out on the web first, with native apps to follow.

How is this different from Perplexity Computer?

Computer ran tasks in temporary sessions that ended with the job. Teammates persist: they keep their role, memory and context, work in parallel, coordinate with each other and with coworkers in Slack, and learn from feedback.

Who else offers something similar?

OpenAI’s workspace agents and Dots, Anthropic’s Claude coworker-style agents, and Microsoft’s Copilot Cowork are all chasing the same ongoing-work model.

Sources: The Verge, Digital Trends, AlphaSignal

Leave a Comment

Your email address will not be published. Required fields are marked *

Get the 5 best AI tools every week

Top AI news, tools, and prompts — one short email. Free, unsubscribe anytime.

Run a newsletter of your own? Monetize and grow it with SparkLoop →

Scroll to Top