Book 1 · Chapter 5: Research and Analysis With AI
This page helps you decide how hard to verify an AI answer based on what is at stake, with a 60-second verification card and the chapter's research prompts.
This page assumes you've read the chapter. It hands you the tools, not the lessons.
Three questions tell you how hard to verify what AI just handed you. The point isn't to second-guess everything: it's to match the checking to what's actually at stake.
Medium stakes(quick spot-check)
Because it's for internal eyes, not an external audience, and being wrong means a bit of rework, nothing worse.
A quick smell test and one fact check is usually enough.
See the 60-Second Verification CardThe goal is to build a proportional verification habit that matches the stakes of the task.
60-Second Verification Card
- Check 1: Does this pass the smell test?
Read the output. Does anything feel off? Claims that seem too specific, too convenient, or too good to be true? Your gut is a decent first filter. If something makes you pause, check it.
- Check 2: Verify the key facts.
Pick the two or three most important factual claims. Open a browser. Spend 30 seconds spot-checking the ones that would embarrass you most if they were wrong.
- Check 3: Check what's missing.
Ask yourself: is there anything obvious that should be here but isn't? Then ask AI directly: “What did you leave out that I should know about?”
From the Three-Check Verification section.
- Focused document summary
I uploaded the annual report from DataVision Analytics, one of our potential vendors. I have a meeting with them Thursday. I need to understand: their revenue trend over the past three years, any mentions of the product line we'd be buying (their Enterprise Analytics Suite), key risks they've disclosed, and whether they mention any major client losses or leadership changes. Keep it under 500 words.
- The “give me a table” comparison
I'm choosing between three project management tools for my team of 20: Monday.com, Asana, and Jira. The things that matter most to us are: ease of use for non-technical team members, pricing for 20 users, integration with Google Workspace, and the ability to create Gantt charts. Give me a comparison table with these four factors and a recommendation based on our priorities.
- The “get me ready” meeting prep
I have a quarterly business review tomorrow with my VP and three department heads. Here are the key metrics from this quarter: - Revenue: $2.1M (target was $2.3M, so we're 9% under) - New clients: 14 (target was 12, so we're ahead) - Client retention: 91% (down from 94% last quarter) - Average deal size: $38K (up from $32K) My VP is going to focus on the revenue miss and the retention drop. I need talking points that: acknowledge the shortfall without spin, explain the context (two large clients churned due to budget cuts, not service issues), highlight the positives (new clients above target, deal size growth), and present a concrete plan for Q3. Also give me three tough questions she might ask and suggested answers.
- Spreadsheet analysis
I uploaded our customer satisfaction survey results from last quarter. 340 responses. I don't know how to analyze survey data. Can you tell me: what's our overall satisfaction score, which areas scored highest and lowest, are there any patterns by department or region, and what are the top three things we should focus on improving? Give me the key findings in plain language that I could present to my team.
- Check what's missing
What did you leave out that I should know about?