Outcome
By the end of this lesson, you'll be able to take an unfamiliar topic — a new tool, a vendor decision, a "wait, what even is this?" question from your boss — and turn it into a clear, decision-ready brief in under an hour. You'll know how to get oriented fast, compare options on a level playing field, and catch the moments where AI sounds confident but isn't actually right.
This is Module B4 in the Work track, right after Inbox, Meetings, and Documents. If those modules got AI handling your busywork, this one gets it handling your thinking — carefully.
Concept
Research and analysis are where AI is most useful and most risky at the same time. Useful, because it can read a hundred pages faster than you can read ten. Risky, because a wrong answer here doesn't just waste your time — it can end up in a client proposal, a budget request, or a decision you're on the hook for.
Remember the hallucination warning from Foundations: AI can produce a confident, well-formatted, completely wrong answer. Nowhere does that matter more than research. A wrong fact buried in a beautifully organized brief is more dangerous than an obviously sloppy guess — it looks trustworthy.
So this module isn't just "how to research faster with AI." It's how to research faster and build in the checks that keep you from acting on something false. Three workflows, one habit running underneath all of them.
The three workflows:
- Orientation brief — get up to speed on something new, fast, with the common misconceptions flagged up front.
- Structured comparison — compare options (tools, vendors, approaches) using the same criteria for each one, not a scattered list of impressions.
- Decision-ready brief — turn all of that into a document that lays out tradeoffs clearly, without pretending AI should make the call for you.
The habit underneath: verify anything that will actually be used in a real decision — before you use it.
Build
1. The orientation brief
When you're new to a topic, don't ask AI to "explain it." Ask it to orient you like someone who already knows the traps.
I'm getting up to speed on [topic] for the first time.
Give me:
1. A plain-language overview (assume I'm smart but new to this)
2. The 2-3 things people commonly get wrong or misunderstand about it
3. The 3-5 terms I'll need to know to follow a conversation about this
4. One question I should ask an expert if I got the chance
Flag anything you're not fully certain about instead of
stating it as fact.
That last line matters. It won't catch every mistake, but it nudges the model toward flagging weak spots instead of papering over them — and gives you a starting list of things to double-check, not a finished answer.
2. The structured comparison
Unstructured pro/con lists feel productive but hide bias — you notice more pros for the option you already liked. Fix this by locking in your criteria before you ask AI to compare anything.
I'm comparing [Option A], [Option B], and [Option C] for [decision context].
Evaluate each one against these criteria:
- [Criterion 1, e.g. cost]
- [Criterion 2, e.g. ease of setup]
- [Criterion 3, e.g. long-term flexibility]
- [Criterion 4, e.g. support/community]
For each option, give a short rating and reasoning per criterion.
Note where you're uncertain or where information may be outdated.
Don't recommend one overall — just lay out the comparison.
Notice the last instruction: don't recommend one overall. That's intentional, and it leads to the next step.
3. The decision-ready brief
This is where scattered research becomes something you'd actually hand to a manager or use to make a call. The move is to ask for tradeoffs, not a verdict — because the decision is yours, and it usually depends on priorities AI doesn't have full visibility into (budget, politics, timing, risk tolerance).
Turn this comparison into a short decision brief for someone who needs to
choose between these options. Structure it as:
1. One-paragraph summary of the situation
2. Key tradeoffs (not a recommendation)
3. What would tip the decision toward each option
4. Facts or figures in here that I should verify independently before
using this in a real decision
That fourth item is the safety net. You're explicitly asking the AI to self-report its own weak points — and then you're going to go check them.
Run
Here's how this fits into a normal week, not just a one-off exercise:
- New topic drops on your desk? Run the orientation brief first, before any meeting or reading. Fifteen minutes buys you enough context to not sound lost.
- Comparing anything with real stakes (a vendor, a tool, a hire, a process change)? Lock your criteria first, always in the same format, so every option gets judged the same way.
- Before anything goes into a real decision — a number, a stat, a claim about what a competitor does — verify it independently. A quick search, a source check, or asking someone who'd actually know. Treat this as a non-negotiable last step, not an optional one.
- When AI hedges or flags uncertainty, don't smooth that over in your final brief. Keep the hedge visible — it's useful information, not noise.
The goal isn't to never trust AI-sourced research. It's to trust it the way you'd trust a smart, fast, occasionally wrong colleague: use their draft, check their facts, and make the call yourself.
Next up in the Work track: turning these briefs into documents and reports that don't just sit in a folder — but that's a module for another day.
Money-related research? This workflow is educational, not financial advice — always verify figures independently before acting on them.