TECH Nov 08, 2026 · by Skills Team

Perplexity Teammates: Always-On AI Agents for Business

Perplexity Teammates are always-on AI agents that own ongoing business work, from research and monitoring to support triage. Learn how they work and how to govern them.

Perplexity Teammates: Always-On AI Agents for Business

The Short Answer

Perplexity Teammates arrived on November 5, 2026, and they change the question businesses ask about AI. Instead of asking what an assistant can answer, teams can now ask what an agent can own. Perplexity Teammates are persistent, always-on AI agents that take responsibility for ongoing work, from competitor monitoring to support triage, and keep working…

Perplexity Teammates arrived on November 5, 2026, and they change the question businesses ask about AI. Instead of asking what an assistant can answer, teams can now ask what an agent can own. Perplexity Teammates are persistent, always-on AI agents that take responsibility for ongoing work, from competitor monitoring to support triage, and keep working until results are ready for review. They run inside Perplexity Computer, remember context between sessions, and work with colleagues through Slack. This guide explains what they are, how they work, and how to govern them.

What Are Perplexity Teammates?

Perplexity Teammates are AI agents built for ongoing work rather than one-off questions. You give a teammate a role, such as research analyst, support triage lead, or QA engineer, and it keeps that role until you change it. It plans the work, breaks a large goal into smaller tasks, runs those tasks in parallel, and follows through until the results are ready for your review. A teammate remembers its goals, instructions, files, and your feedback across sessions, unlike a chatbot session that ends when you close the window.

Each teammate runs on its own isolated computer inside Perplexity Computer, with its own desktop, browser, and connected tools. That separation lets several teammates work at the same time without stepping on each other. A marketing teammate can track competitors while a sales teammate prepares outreach and an operations teammate chases a missing deliverable. Teammates can also hand work to one another, so a research agent can pass findings to an engineering agent without a human acting as messenger.

Chief executive Aravind Srinivas announced the launch on X on November 5, 2026, saying teammates work for you and envisaging hundreds of teammates per person. Perplexity reports its staff built around 10,000 teammates in the 30 days before launch. The feature is a research preview, so it is live and usable but still being developed.

  • Persistent roles: a teammate keeps its job until you reassign it.
  • Parallel work: a teammate can start extra agents at the same time, each with its own computer.
  • Shared memory: goals, files, skills, and feedback carry across sessions.
  • Team presence: teammates can join Slack channels, where colleagues assign work or ask for updates.
  • Agent handoffs: teammates pass tasks between themselves when a job needs more than one skill set.

How Perplexity Teammates Work

The workflow starts with a high-level goal, not a detailed prompt. Describe the outcome you want, for example a weekly competitor briefing or a triaged support inbox, and the teammate plans the steps, gathers information, and carries out the work. It draws on your Perplexity Computer projects, skills, memory, files, and connected applications, so it starts with context instead of a blank slate, and notifies you when it needs a decision.

Monitoring is built into the design. Every teammate has its own dedicated thread inside the Perplexity app, where you can watch progress, review what it has done, give extra instructions, or take over yourself. This is the control panel for always-on work, and it is what separates Teammates from a fire-and-forget automation.

The Slack integration is the other half of the story. Teammates can be added to Slack channels as apps, so anyone on the team can hand work to an agent or ask how a job is going without opening another tool. A teammate can also watch a channel and act when something changes, which suits monitoring work such as new support tickets or competitor mentions. New teammates start on Claude Opus 5.5, changeable per teammate or task.

Perplexity Teammates vs OpenAI Dots and Meta Muse

Perplexity is not alone in this shift. OpenAI has launched Dots, its own always-on agents for business, and Meta is pushing in the same direction with Muse, its agent platform for workplace use. All three share the same idea: AI that holds an ongoing job instead of answering a single question. The differences lie in where the agents live and how teams reach them.

Our guide to OpenAI Dots explained covers OpenAI's always-on agents in detail. In brief, Dots brings persistent agents into the ChatGPT ecosystem, while Teammates brings them into Perplexity Computer and Slack, with each agent working on its own virtual computer. Meta Muse adds a voice-led angle to agent work, which suits hands-free use cases. For most businesses, the deciding factor is the tools your team already uses every day.

FeaturePerplexity TeammatesOpenAI DotsMeta Muse
Core ideaPersistent agents that own ongoing workAlways-on agents for business teamsAgents for workplace use, with a voice-led experience
Where teams reach itPerplexity web app, plus Slack, GitHub, and LinearChatGPT ecosystemMeta's agent platform and voice interfaces
Parallel workYes, each teammate runs on its own computerYes, agents run ongoing tasks in the backgroundBuilt for continuous, hands-free assistance

These tools will likely coexist in most businesses, so what matters is building the governance habits that let any always-on agent work safely, which we cover below.

Business Use Cases for Perplexity Teammates

Perplexity points to five business functions where Teammates can help: marketing, sales, support, operations, and finance. Here is what that looks like in practice.

Research and Monitoring

This is the strongest starting point. A research teammate can track competitors, watch for industry news, monitor regulatory changes, and deliver a short briefing on a schedule you set.

Marketing and Sales

A marketing teammate can draft campaign ideas, check ad performance, watch competitor messaging, and prepare content briefs. A sales teammate can find stalled conversations in your pipeline, prepare outreach drafts, and summarise what a prospect needs before a call. These repeatable jobs gain real capacity from an agent that never sleeps, as long as a human reviews customer-facing output.

Support and Operations

A support teammate can summarise yesterday's complaints, draft replies for your team to approve, and spot patterns in recurring issues. An operations teammate can chase missing items before a deadline and coordinate handoffs between departments. Software teams can use engineer agents to investigate alerts and QA agents to run testing loops.

These use cases connect directly to wider business process automation, where the goal is to remove the repetitive work that slows people down. If you are mapping out which processes to automate first, our complete guide to business process automation in 2026 walks through the same decision in detail.

What Always-On Agents Mean for Teams

Always-on agents change how a team organises its day. When an agent owns a recurring job, the human role shifts from doing the task to defining it well and reviewing the result. Goals must be written clearly, because an agent works from the goal you gave it. Success criteria must be explicit, because an agent cannot guess what good looks like. And review must be scheduled, because an agent that runs while you sleep can also drift while you sleep.

The biggest gain is continuity. A human researcher goes home, takes leave, and forgets details. A teammate keeps the thread, remembers the context, and picks up where it left off. Treat the first month as training, with a human reviewing everything.

Governance and Oversight for Perplexity Teammates

Perplexity Teammates make governance urgent, because an agent that works while you are away can also act while you are away. The same features that make Teammates useful, dedicated threads, review steps, and Slack visibility, also make them governable. The key is to set the rules before you scale, not after something goes wrong.

Start by limiting what each teammate can touch. Give an agent access only to the tools, files, and channels its role genuinely needs. A monitoring agent does not need finance system access, and a drafting agent does not need permission to message customers. Keep humans in the loop for anything external or irreversible: drafts are fine for an agent to prepare, but sending to a client, publishing, or changing a live system should wait for human approval.

As agent numbers grow, governance must grow with them. Perplexity's chief executive talks about hundreds of teammates per person, and at that scale no team can manage agents one by one. Our guide to governing hundreds of AI agents explains how to build an agent inventory and audit what agents can access. Pair that with a secure AI adoption playbook for business teams so access and approval rules are written down first.

  • One owner per agent: every teammate should have a named human responsible for its work and access.
  • Least privilege access: connect only the tools and data a role needs, and review those connections regularly.
  • Human approval for external actions: nothing goes to a customer, a publisher, or a live system without a person signing off.
  • Regular thread reviews: check each teammate's thread on a set schedule and correct drift early.
  • Clear off switch: know how to pause or remove a teammate quickly if its behaviour changes.
  • Data boundaries: decide in advance which files and systems agents may never touch, especially customer data.

Getting Started with Always-On Agents

Start small and prove the value before you expand. A careful pilot takes weeks, not months.

  1. Pick one recurring job: choose a task that repeats weekly, takes real time, and has a clear output, such as a competitor briefing or a support summary.
  2. Write the goal down: define the outcome, the sources to use, the format, and what good looks like, then give that brief to the teammate.
  3. Run it in review mode: for the first few weeks, review every output in the teammate's thread and give specific corrections.
  4. Connect it to your team: once the output is reliable, add the teammate to the relevant Slack channel so colleagues can use it too.
  5. Scale one role at a time: add a second teammate only when the first is producing reviewed, trusted work on schedule.

For many businesses, Teammates will be one part of a wider automation plan. Mapping repetitive work across marketing, sales, support, and operations first will show which jobs suit an always-on agent.

Frequently Asked Questions

What are Perplexity Teammates?

Perplexity Teammates are persistent, always-on AI agents that take ownership of ongoing business work instead of answering one-off questions. Each teammate holds a role, plans its own tasks, and runs them in parallel on its own computer until results are ready for review.

When did Perplexity launch Teammates?

Perplexity announced Teammates on November 5, 2026, with its launch blog post dated November 4, 2026. The feature arrived as a research preview for individual Max subscribers on the web, with iOS and native apps planned for later.

How are Perplexity Teammates different from a normal AI chatbot?

A chatbot waits for your next prompt and starts fresh each time. A teammate keeps its role, remembers goals, files, and feedback between sessions, monitors tools such as Slack, GitHub, and Linear, and acts when something changes.

Can Perplexity Teammates work inside Slack?

Yes. Teammates integrate with Slack, so you can add them to channels where colleagues can assign tasks, ask for updates, or check progress. A teammate can also watch a channel and act when something relevant appears.

How do I monitor what a teammate is doing?

Every teammate has its own dedicated thread in the Perplexity app, where you can follow progress, review completed work, add instructions, or take over the task yourself. Slack updates add a second way to stay in the loop.

Are Perplexity Teammates ready for business use today?

They are usable today as a research preview, but businesses should treat them as an early release. Start with a small pilot, keep a human reviewing every output, and expand only once results are consistently reliable.

Always-on agents are changing how teams handle research, monitoring, and reporting. The winners will combine these tools with clear goals and strong oversight. If you want help mapping which of your processes suit an always-on agent, or building the automation around it, talk to our team and we will plan the right starting point for your business.

#Perplexity Teammates#AI agents#business automation#always-on AI