I've spent the last few months going deep on Perplexity AI — testing every plan, every feature, every workflow I could think of.

Here's what I found: most people sign up, ask a few questions, and never touch the settings. They're leaving 80% of the value on the table.

Here are 5 quick ones:

1. Default Focus Mode — Most people leave it on "All." But if you do research-heavy work, switching your default to "Academic" or "Writing" changes the quality of responses dramatically.

2. AI Model Selection — Perplexity lets you choose which model powers your answers. GPT-4o, Claude, Sonar — each has strengths. Matching the model to your task type is a game-changer.

3. Collection Organization — Stop letting your threads pile up. Collections let you organize research by project. Think of them like folders for your AI brain.

4. Pro Search vs Quick Search — Pro Search does multi-step reasoning and follow-up queries automatically. Most people don't realize when to use it vs. quick search. Wrong choice = worse answers.

5. Output Formatting — You can guide Perplexity's output format with simple instructions in your query. Tables, bullet points, comparisons — it's all possible if you ask right.

These are just the basics. I put together a full setup guide covering plan selection, every feature config, and workflow templates.

If you want the complete playbook → check the link in my bio.