Day 1
Answer engine, not chatbot
Understand what Perplexity actually does differently and when it beats a normal chatbot or a Google search.
Concepts
- · Perplexity retrieves live web results first, then writes an answer grounded in them
- · Every claim carries a numbered citation you can open and verify
- · Good for: current facts, comparisons, research triage. Weak for: long creative writing, private data reasoning
- · Free vs Pro: free gives limited Pro searches per day, Pro unlocks more searches, model choice and file uploads
Exercise
Ask the same question three ways — in Perplexity, in a plain chatbot, and in Google — and write down which gave you a decision you'd actually act on.
Try this prompt
What changed in [your industry] in the last 90 days that affects [your specific job]? Cite primary sources and note publication dates.
Mastery signal: You can explain in one sentence when you'd reach for Perplexity instead of ChatGPT or Google.
Day 2
Reading the sources like a pro
Stop trusting the summary and start using the citation trail — this is where the real value is.
Concepts
- · Open the source panel and check publish dates, not just domains
- · Spotting when Perplexity paraphrased a low-quality aggregator instead of the primary source
- · Asking follow-ups that force better sourcing ('use only official documentation from the last 12 months')
- · When an answer has no citation on a claim, treat that sentence as unverified
Exercise
Run a question in your own field where you already know the correct answer. Audit every citation and count how many were primary sources.
Try this prompt
Answer using only primary sources — official docs, filings, or peer-reviewed papers published after [date]. If a claim has no such source, say so explicitly instead of guessing.
Mastery signal: You caught at least one weak or outdated citation and corrected the answer with a follow-up.
Day 3
Search modes and model choice
Match the mode to the task instead of using the default for everything.
Concepts
- · Quick/Search for fast factual lookups
- · Pro Search for multi-step questions that need several rounds of retrieval
- · Research (Deep Research) for long-form reports — slower, far more sources
- · Focus filters (Academic, Social, Finance, YouTube) to constrain where it looks
- · Model selection on paid tiers: reasoning models for analysis, faster models for lookups
Exercise
Take one real work question and run it in Quick, Pro, and Research mode. Time each and note where extra depth actually changed your conclusion.
Try this prompt
Compare [option A] and [option B] for [your specific constraint: budget, team size, regulation]. Give a recommendation, the strongest counter-argument, and what evidence would change your mind.
Mastery signal: You can predict, before running a query, which mode it needs — and you're right most of the time.
Day 4
Prompting for retrieval, not conversation
Write queries that steer what Perplexity searches for, not just what it says back.
Concepts
- · Include time bounds, geography, and source type in the query itself
- · Name the output format up front: table, pros/cons, checklist, decision memo
- · Give your constraints (role, budget, stack, regulation) so the answer is scoped to you
- · Iterate in-thread — follow-ups reuse the retrieved context instead of starting over
Exercise
Rewrite your three most common work questions as scoped prompts with time bounds, format, and constraints. Save them somewhere reusable.
Try this prompt
I'm a [role] at a [size] company in [industry], operating under [constraint]. Give me a comparison table of the top 5 options for [task], with cost, setup effort, and the main risk of each. Sources must be from 2025 or later.
Mastery signal: Your first query usually returns something usable, without three rounds of clarification.
Day 5
Spaces, files and your own context
Turn Perplexity from a one-off search box into a place your recurring work lives.
Concepts
- · Spaces group threads, files, and custom instructions around one project or client
- · Custom instructions per Space set tone, role context, and answer format once
- · Uploading PDFs, CSVs and docs lets you ask questions across your own material plus the web
- · Files stay in the Space — check your organization's policy before uploading anything sensitive
Exercise
Create one Space for a real ongoing project. Add custom instructions describing your role and preferred output, upload two relevant documents, and ask a question that needs both the files and the web.
Try this prompt
Using the uploaded documents plus current public information, identify where our approach is now out of date and what specifically I should change first.
Mastery signal: You've stopped re-explaining your context in every prompt because the Space already holds it.
Day 6
Labs, Comet and the newer surfaces
Know what the advanced tiers actually add — and whether they're worth your subscription.
Concepts
- · Labs turns a prompt into a deliverable: a small report, dashboard, spreadsheet or simple web app
- · Comet, Perplexity's browser, brings the assistant onto pages you're already reading
- · Mobile and voice for capture on the go; browser extension for in-context questions
- · Honest limits: generated apps and reports need review, and heavy Labs use burns through quota fast
Exercise
Pick a deliverable you'd normally build by hand — a competitor grid, a briefing doc — and have Labs draft it. Then log how long editing it took versus starting from scratch.
Try this prompt
Build a one-page briefing on [topic] for [audience]: current state, three developments that matter, what we should do, and a source list with dates.
Mastery signal: You can say, with a number, whether your current tier pays for itself.
Day 7
Making it a habit and staying unbiased
Fit Perplexity into your actual week and build the verification reflexes that keep you safe.
Concepts
- · Three fixed slots beat random usage: morning scan, mid-task research, end-of-week synthesis
- · Always verify before anything leaves your desk: numbers, names, dates, quotes, legal or medical claims
- · Perplexity has commercial partnerships and ranks sources — cross-check anything high-stakes elsewhere
- · Know when to switch tools: private data, long drafting, or code work often belong somewhere else
Exercise
Write a five-line personal usage policy: what you'll always use it for, what you'll never use it for, and what you always double-check.
Try this prompt
Here's my draft conclusion: [paste]. Argue the strongest case against it using current sources, and tell me which of my assumptions is weakest.
Mastery signal: You've used it on three real tasks this week and can name one thing you deliberately did not use it for.