Skip to content
Gab AI

How to Use Perplexity AI Teammates to Automate Your Daily Workflow on Mac

Perplexity AI Teammates just landed on Mac. Here's how freelancers, creators, and small business owners can set up persistent AI agents for research, drafting, and daily task automation.

Arya · · · 16 min read

How to Use Perplexity AI Teammates to Automate Your Daily Workflow on Mac

You know that moment around 2 PM when you realize you've spent the entire morning bouncing between tabs — researching a competitor, drafting a newsletter intro, scanning your inbox for anything urgent, checking if that blog post idea you had last week is even worth pursuing? You haven't done any real work yet. You've just been gathering ingredients.

That's the problem Perplexity AI Teammates is designed to solve. And now that it's available on Mac desktop as of November 14, 2026, it opens the door for the kind of sustained, professional work that freelancers, creators, and small business owners actually do all day.

This isn't a gimmick. It's a shift in how you can structure your workday — if you set it up thoughtfully. Let me walk you through exactly how to do that.

What Are Perplexity AI Teammates, Really?

Forget the marketing language for a second. Here's the core concept: Perplexity AI Teammates are persistent AI agents that you configure once and then run repeatedly. Think of them less like a chatbot you talk to and more like a junior assistant who shows up every morning already knowing what you need.

Each "teammate" is essentially a saved configuration — a set of instructions, a defined role, and a task scope. You might have one teammate that monitors industry news every morning, another that drafts content briefs based on your editorial calendar, and a third that summarizes long research papers into bullet points you can actually use.

The key word here is persistent. Before features like this, you'd open a new AI chat, re-explain your context, re-state your preferences, and hope the output matched what you needed. With a persistent agent approach, that setup is done once. The agent remembers its role, its constraints, and your preferences across sessions.

The Mac desktop rollout is significant because it means you can now run these agents alongside your actual work environment: your writing app, your browser, your project management tool, your email. That's when persistent AI agents stop being a novelty and start being genuinely useful. Mobile and web interfaces have their place, but desktop availability is what makes sustained professional workflows practical.

A note on access: As of this writing, Perplexity offers both free and paid tiers. The exact subscription tier required for AI Teammates has not been publicly confirmed in detail, so check Perplexity's current pricing page before assuming access. Features like this are typically gated behind a paid plan, but verify before subscribing.

Who This Is Actually For (And Who Should Wait)

Let me be honest: if you use AI once a week to brainstorm blog titles, you don't need this. Teammates is built for people who interact with AI multiple times per day and are tired of repeating themselves.

The sweet spot is:

  • Freelancers juggling 3-5 clients who each need research, drafts, and status updates
  • Content creators who publish regularly and need a steady pipeline of ideas, outlines, and first drafts
  • Small business owners who handle their own marketing, customer research, and competitive analysis
  • Church leaders or nonprofit directors who write weekly communications, plan events, and need to stay informed on community issues

If you're already using AI for research and writing, Teammates is the difference between having a tool and having a system. If you're brand new to AI, start simpler — get comfortable with basic prompting first, then come back here when you're ready to scale up.

Setting Up Your First AI Teammate on Mac: Step by Step

Here's the practical walkthrough. I'm assuming you have a Perplexity paid subscription (check their pricing page for the current tier that includes Teammates) and have the Mac desktop app installed.

Important caveat: The exact interface and setup fields may vary as Perplexity updates the feature. The steps below reflect the general workflow for configuring a persistent AI agent. If the specific labels or options in your app differ from what's described here, the underlying principles still apply — you're defining a role, setting instructions, and specifying output preferences. Adapt as needed based on what you see in the app.

Step 1: Identify Your Most Repetitive AI Task

Before you touch the app, answer this question: What do I ask AI to do at least three times a week?

Don't start with something ambitious like "manage my entire content strategy." Start with something boring and repetitive. The best first teammate handles a task you already know how to do — you just don't want to keep doing it manually.

Good first teammates:

  • Morning industry news summary
  • Weekly competitor blog scan
  • Meeting notes cleanup and action item extraction
  • First-draft email responses for common inquiries

Bad first teammates:

  • "Be my CMO" (too vague)
  • "Handle all my client communication" (too risky without review)
  • "Write my entire blog" (too much trust, too soon)

Step 2: Create the Teammate Profile

Open the Perplexity Mac app and navigate to the Teammates section. When creating a new teammate, you'll generally want to define the following elements — whether the app presents them as separate fields or as part of a single configuration prompt:

  1. Name and role — Give it a clear, functional name. "Morning Research Brief" is better than "Research Bot." The name should remind you what it does at a glance.
  2. Core instructions — This is where you describe the teammate's job. Be specific about scope, tone, and output format. (I'll give you copy-paste prompts below.)
  3. Source preferences — If you want the teammate to prioritize certain publications, industry sites, or data sources, specify them in your instructions.
  4. Output format — Do you want bullet points? A narrative summary? A table? Define this upfront so you don't have to reformat every time.
  5. Review expectations — Decide whether you'll use this teammate daily, weekly, or on-demand, and note that in your instructions so the agent calibrates its depth accordingly.

Step 3: Write the System Prompt (This Is Where Most People Fail)

The system prompt — the core instructions you give your teammate — is the single most important part of your setup. It's the difference between getting useful output and getting generic AI filler.

Here's the framework I recommend:

Goal → Audience → Context → Constraints → Format → Review trigger

Let me show you what this looks like in practice.

Copy-Paste Prompts for Common Workflows

These are ready to use. Paste them into the instructions field when creating a teammate, then adjust the bracketed sections for your situation. These prompts follow universal agent-prompting principles — they'll work in Perplexity AI Teammates and can be adapted for other persistent AI agent tools as well.

Prompt 1: Morning Competitive Research Brief

Text example
Text
You are my competitive research analyst. Every morning, your job is to scan for new developments from [list 3-5 competitors or industry players].

Focus on: product launches, pricing changes, new content published, hiring signals, and partnership announcements.

Ignore: routine social media posts, minor UI updates, or anything older than 7 days.

Format: A bulleted summary with no more than 5 items. Each item should be 2-3 sentences: what happened, why it matters, and whether I need to respond.

Tone: Direct and analytical. No hype. Flag anything that directly affects [your business/niche].

If nothing meaningful happened, say so in one sentence. Don't pad the report.

How to adapt this: The bracketed sections are where you make it yours. List your actual competitors by name. Define your niche specifically — "B2B SaaS for accounting firms" is better than "tech." If you care more about content moves than hiring signals, reorder the focus list. After your first few runs, you'll notice which categories consistently produce noise — remove them.

Prompt 2: Content Brief Generator

Text example
Text
You are my content strategist. When I give you a topic or keyword, produce a content brief that includes:

1. Working title (under 60 characters)
2. Target audience (be specific — not "everyone")
3. Search intent (what is someone Googling this actually trying to do?)
4. 3-5 key points the article must cover
5. 2-3 angles that most existing articles miss
6. Suggested structure (H2 headings)
7. One concrete example or case study to include

Do not write the article. Just the brief.

Context: I write for [describe your audience]. My tone is [describe your voice]. I publish on [your platform].

Constraint: The brief should fit on one page. If a section doesn't add value, skip it.

How to adapt this: The "angles that most existing articles miss" instruction is what makes this prompt valuable — it pushes the agent past generic advice. But you need to give it enough context about your audience to find genuinely fresh angles. If you write for experienced developers, say so — otherwise you'll get beginner-level suggestions. Update the context section every quarter as your audience evolves.

Prompt 3: Inbox Triage Assistant

Text example
Text
You are my email triage assistant. When I paste a batch of email subjects and senders (or summaries), categorize them into:

1. **Respond today** — time-sensitive, from important contacts, or revenue-related
2. **Respond this week** — important but not urgent
3. **Delegate or template** — can be handled with a standard reply
4. **Archive** — newsletters, notifications, or FYI-only messages

For each "Respond today" item, draft a 2-3 sentence reply suggestion.

Context: I am a [your role]. My priorities are [list 2-3 current priorities]. Important contacts include [list key names or organizations].

Constraint: Never auto-send anything. All replies are drafts for my review.

How to adapt this: The priority list is critical. Update it weekly — what's urgent during a product launch is different from what's urgent during a quiet planning week. The "important contacts" list should include anyone whose email you'd regret missing for 24 hours. And that last constraint — "never auto-send" — is non-negotiable. Keep it.

Notice what these prompts have in common: they're specific about what to include and what to leave out. They define format. They set boundaries. That last part — constraints — is what separates a useful AI teammate from one that generates noise.

If you want to explore building custom AI agents for similar workflows across different tools, the underlying principles are the same: clear role, defined scope, explicit output format.

Designing a Daily Workflow With Multiple Teammates

Once you've got one teammate working well, the real power comes from chaining them into a daily rhythm. Here's an example workflow for a freelance content creator. Note that this workflow assumes you're manually triggering each teammate at the appropriate time — whether Perplexity supports automated scheduling for Teammates is not confirmed, so plan on initiating each session yourself.

7:00 AM — Morning Brief Teammate scans your industry and delivers a 5-item summary. You read it with coffee. Takes 3 minutes.

9:00 AM — Content Brief Teammate takes the topic you picked from yesterday's brainstorm and generates a structured brief. You review and adjust for 10 minutes, then start writing.

12:00 PM — Research Teammate pulls supporting data, statistics, or expert quotes for the piece you're drafting. You review sources for accuracy.

3:00 PM — Inbox Triage Teammate categorizes the emails that piled up while you were doing deep work. You spend 20 minutes on responses instead of 60.

5:00 PM — End-of-Day Summary Teammate compiles what you accomplished and flags anything unfinished for tomorrow.

The total AI interaction time here is maybe 45 minutes. But it replaces 2-3 hours of scattered, low-value task-switching. That's the real win — not that AI does your work, but that it handles the preparation so you can focus on the work that actually requires your brain.

This kind of daily rhythm works regardless of which AI agent tool you use. The principle is the same: batch your AI-assisted tasks into predictable time blocks so they support your deep work instead of interrupting it.

What Most People Get Wrong With AI Teammates

I've seen a few patterns that consistently lead to disappointing results. Avoid these.

Mistake 1: Making the Scope Too Broad

"Be my marketing assistant" is not a useful instruction. It's like hiring someone and saying "do marketing." The teammate doesn't know if you want social media copy, email campaigns, ad analysis, or brand strategy. Pick one task per teammate. You can always create more.

Mistake 2: Skipping the Constraints

If you don't tell the teammate what to exclude, it'll include everything. Your morning brief will be 2,000 words of marginally relevant noise. Your content briefs will suggest angles for audiences you don't serve. Always define what's out of scope.

Mistake 3: Trusting Output Without Review

This is the big one. AI teammates are drafters, not decision-makers. Every output should pass through your judgment before it goes anywhere — especially anything client-facing, public, or financial.

Specific things to always verify manually:

  • Statistics and data points (AI still fabricates numbers)
  • Quotes and attributions
  • Competitor claims (is that pricing change real, or hallucinated?)
  • Tone and appropriateness for your specific audience
  • Legal or compliance-sensitive language

Build a 60-second review habit for each teammate's output. Scan for anything that feels off. Check one fact per report. It doesn't have to be exhaustive — it has to be consistent.

Mistake 4: Never Updating the Prompt

Your first prompt won't be perfect. After a week, you'll notice patterns: the teammate over-explains certain things, misses a source you care about, or formats output in a way that doesn't match your workflow. Update the instructions. A good AI teammate setup is a living document, not a set-and-forget configuration.

Here's a simple review cadence: after your first three runs, make one adjustment. After the first full week, do a more thorough review. Then check in monthly. Keep a running note of what you changed and why — it'll help you set up future teammates faster.

Mistake 5: Using AI Teammates for Tasks That Need Human Judgment

Some things shouldn't be delegated to AI, even as drafts. Sensitive client communications, creative direction decisions, pricing strategy, and anything involving personal relationships — these require your instincts, your context, and your accountability. AI can prepare you for these moments. It shouldn't handle them.

A useful litmus test: if the output going wrong would damage a relationship or cost you money, a human needs to own it from the start — not just review it at the end.

When AI Teammates Aren't Enough (And What to Do Instead)

Perplexity AI Teammates are good at a specific thing: research-backed text tasks with clear parameters. But your workflow probably includes more than that.

You might need to generate images for a blog post, create a short video for social media, draft music for a podcast intro, or build a presentation. Teammates won't do that — they're text-focused agents tied to Perplexity's research engine.

This is where having an all-in-one AI tool matters. Instead of subscribing to one service for research agents, another for image generation, another for video, and another for writing — you can consolidate. Gab AI lets you create text, images, videos, music, and more from a single dashboard. For freelancers and small business owners already juggling too many subscriptions, that consolidation isn't just convenient — it's a meaningful reduction in cost and friction.

If you're building out a research-heavy workflow specifically, the research assistant tool is worth exploring as part of that consolidated approach.

Quick Start: Your First 10 Minutes

If you want to get started right now, here's the minimum viable setup:

  1. Pick your most repetitive research task — the one you do at least 3x per week.
  2. Open the Perplexity Mac app and create a new teammate.
  3. Paste the competitive research prompt above (or whichever matches your task). Fill in the brackets.
  4. Run it once and evaluate the output honestly. Is it useful? What's missing? What's noise?
  5. Adjust the prompt based on that first run. Add one constraint and one formatting tweak.
  6. Run it again tomorrow. Compare. Adjust again if needed.
  7. After three good runs, consider adding a second teammate for a different task.

That's it. Don't try to build a five-agent system on day one. Build one that works, then expand.

For more step-by-step AI workflows like this, browse the guides and how-to section — there's a growing library of practical walkthroughs for different use cases.

Self-Check: Is Your Teammate Setup Working?

After your first week, run through this quick checklist:

  • Does the teammate's output save you time compared to doing the task manually?
  • Are you actually using the output, or just generating it and ignoring it?
  • Have you updated the prompt at least once based on real results?
  • Are you reviewing every output before acting on it?
  • Is the scope narrow enough that the output is consistently relevant?
  • Have you caught at least one factual error during review? (If not, you may not be reviewing carefully enough.)

If you answer "no" to more than two of these, simplify. Narrow the scope, tighten the constraints, or switch to a different task entirely. A teammate that generates output you don't trust or don't use is worse than no teammate at all — it's a distraction disguised as productivity.

The Bigger Picture: AI Agents Are Becoming AI Systems

What Perplexity is doing with Teammates on Mac is part of a broader shift. AI tools are moving from one-off conversations to persistent, role-based systems that integrate into your daily routine. This is happening across the industry — not just with research tools, but with writing, creative work, and business operations.

The people who benefit most won't be the ones with the fanciest AI setup. They'll be the ones who took time to define what they actually need, built simple systems around those needs, and maintained the discipline to review AI output before acting on it.

AI teammates don't replace your judgment. They free it up for the work that matters.

That's a meaningful difference — if you set it up right.


Ready to go beyond research agents and bring text, images, video, and music generation into one workflow? Start creating at Gab AI — or download the app and see what a unified AI dashboard feels like in practice.

Back to Blog

Related Articles

Ready to try Gab AI?

One AI for everything — text, images, videos, music, voice, and more. Start free today.