In my last post on AI browsers, I explored how tools like Comet and Dia were transforming the browser from a passive window into an intelligent partner — capable of summarizing, scheduling, and searching smarter.
That was the why.
This one’s the how.
Because AI browsers just crossed a line.
They no longer just find information — they act on it.
And the best example right now is Perplexity Comet, which quietly turned the humble browser into an agentic automation engine.
In today’s Playbook, I’ll show you step-by-step, how I built ten mini agents inside Comet that now handle research, prospecting, trend tracking, and meeting prep automatically.
If you’ve been wondering what “AI in your workflow” actually looks like in practice, this is it.
🌐 From Search to Execution
Most AI tools answer questions.
Comet executes workflows.
You can use it in two ways:
🧠 Main Chat Window — ideal for complex, multi-step tasks.
Think: “Analyze all competitor pages and export a full report.”
⚡ Sidebar Assistant — perfect for contextual actions.
Think: “Summarize this article while I’m reading it” or “Draft an email based on this page.”
This dual setup lets you move between contextual assistance and autonomous execution — without ever leaving the browser.
Now, let’s explore what that looks like in real work.
🔍 1. Live Competitor Research Agent
Traditional competitor research means juggling tabs, spreadsheets, and screenshots.
Comet fixes that.
I open my product site and my top 3–4 competitor pages, then type:
“Analyze all open tabs and summarize the differences in pricing, messaging, and features. Include screenshots and export to Google Docs.”

Here’s what happens behind the scenes:
Comet reads each tab, not just the visible content.
It captures screenshots for reference.
It creates a structured report — usually with a comparison table and highlights.
And it can export directly to Google Docs or Notion if
connected.

What took hours before now takes about 15 minutes — and it’s fully repeatable.
This is the moment you realize your browser isn’t just for browsing anymore.
🧠 2. Research Agent + Organized Workspaces
Research is where Comet shines.
Instead of 20 cluttered tabs, you can ask it to group related pages into a dedicated workspace — “AI Marketing,” “Crypto Regulation Research,” or whatever project you’re on.

Then, the magic:
“Read all open tabs in this workspace. Identify 3–5 key themes, main insights, and implications.”
Comet scans every tab, finds overlapping ideas, and produces a coherent summary.
No note-taking. No context loss.

You can also:
Save the command as a shortcut (
/summarize_workspace)Re-run it anytime with a single prompt
Choose which model or search mode it uses (basic, deep, or fact-check)

And speaking of fact-checking — Comet does that too.
Highlight a claim on a page → right-click → “Verify with Perplexity”.
It cross-references live sources and gives you a quick credibility check.
For anyone doing research, analysis, or writing — this is an enormous time-saver.
🗞️ 3. Trend Tracker & News Synthesizer
Once I saw how easily Comet pulled research, I extended it into automation.
I created a Notion database called Weekly AI SEO News.
Then I asked Comet to:
“Find the top 5 AI SEO stories from the past week and update them into my Notion table with title, date, summary, and link.”

Within minutes, it:
Searched the web in real-time
Compiled relevant headlines
Wrote short summaries
And automatically inserted them into Notion
You can even schedule it weekly, so it runs on its own.

Think of it as your news analyst agent, constantly curating what matters most to your projects.
And this isn’t limited to news — it works for any kind of recurring research update (market trends, tech releases, regulatory updates, etc.).


