A free preview of the habits that make any AI tool more useful — not a tour of features, just what actually changes how you work. The full depth lives in our role-specific modules.
Two minutes of setup, and only one thing is actually required.
Pick one tool and stick with it for now. Which one matters far less than actually using it — you can compare others later, in Lesson 1.
Open a new chat. That's the whole setup. Every lesson below works inside whichever tool you picked.
No prior experience assumed. If you've never used any of these tools, or only played with one once and stopped, this course is built for exactly that starting point.
Three tools, endless opinions about which is "best," and no obvious way to know where to start.
Rough positioning that's good enough to start with: ChatGPT is the strongest general all-rounder for writing, brainstorming, and quick analysis. Claude tends to be stronger with longer documents and careful, detailed instructions. Perplexity is built for research-style questions and shows its sources as it answers. Pick one based on what you'll actually use it for most, not which one is trending.
Ask your chosen tool a real question you had this week — something you'd normally Google or ask a colleague. See how it responds before judging the tool on anything else.
Trying to master all three at once. One tool used well beats three tools used badly — you can always add a second later once the first is a habit.
A vague request gets a vague answer, and it's easy to blame the tool when the real issue is the question.
Be specific about what you want, who it's for, and what "good" looks like. The more context you give upfront, the less back-and-forth you need — and the less generic the answer feels.
Take a vague request you'd normally type — "write me something good" — and rewrite it with three specifics added: who it's for, the tone you want, and one example of what a good result would look like.
Judging a tool by its answer to a vague question. The fix is almost always a better question, not a different tool.
Describing a situation from memory instead of pasting the actual text or data leads to AI inventing details that were never there.
Whenever you have the real source in front of you — an email, a document, a table — paste it directly rather than summarizing it yourself first. Every time you paraphrase before asking, you lose detail the AI can't get back.
Paste an actual email or message below, then ask: summarize this in two sentences, without adding anything that isn't in the text.
Summarizing something yourself first and then asking AI to work from your summary. You lose detail at every hand-off — go to the source directly.
Confident-sounding answers aren't always correct — especially with dates, names, numbers, or anything niche.
Treat AI like a fast first draft, not a finished answer. Verify anything specific — a stat, a quote, a name, a date — against a real source before you use it anywhere that matters.
Take an answer AI gave you earlier and ask: list every specific fact, number, or claim in this answer, so I can check each one.
Trusting an answer more because it's confidently worded. Confidence and accuracy aren't the same thing — AI sounds equally sure whether it's right or wrong.
It's tempting to let AI make calls it shouldn't — personal decisions, sensitive judgment calls, anything with real consequences for someone.
Use AI to prepare options and lay out information — keep the actual judgment call with a person, especially anything involving another real person's data, wellbeing, or circumstances.
Ask for the pros and cons of a real decision you're weighing, but explicitly say: don't tell me what to choose, just lay out the considerations.
Treating an AI's suggestion as a decision already made, rather than as one input among several. It's a fast way to think something through, not a substitute for thinking it through.
General information about your day, your work, or a project is fine to paste. Anything with another real person's private details — their contact information, health information, or anything under an NDA — isn't, regardless of which tool or which pricing tier you're on. When in doubt, check your own company's data policy rather than assuming a tool's terms cover you.
These aren't specific to one role — they're the foundation everything else builds on, including our role-specific modules.
Now that you've got the fundamentals, the role-specific modules go much further — built around your actual job, not general habits.
