Not sure where AI fits in your business? Let it interview you about last week

How to do it

1. 1

   Paste the interview prompt and answer its questions

   It asks one question at a time, 15 at most, and asks for counts and minutes when you're vague. The whole thing takes about 45 minutes.

   Prompt

   Expand ⌄Copy

   I run an established business. I'm new to AI. I don't want generic ideas. I want to find where AI can save my team real time in the next 90 days, starting with one small pilot. Act as a practical, hype-free operations advisor. First, interview me: - Ask one question at a time. Wait for my answer. Ask 15 questions at most. - Ask about what happened last week, not what "usually" happens. - Cover: what we sell and to whom, team size, the software we use, documents and emails we write again and again, data someone copies from one place to another, questions people keep asking, work that waits on one person, tasks we never get to, and mistakes that cost us money. - If I'm vague, ask for numbers: how often, how many minutes, who. - Don't suggest anything until the interview is done. Then give me: 1. A table of every repeated task: task | who | how often | hours per month. 2. The top 3 AI opportunities, ranked. For each: hours saved per month (be conservative, show the math), the time a person needs to check the AI's work, what happens if the AI is wrong, and who should own it (not me). 3. ONE pilot my team can start this week with a normal AI chat tool. No new software. No IT project. Nothing reaches a customer without a person checking it. 4. A "not yet" list: ideas that sound good but should wait, and why. 5. A 2-week test plan: what to measure first, what the owner does each day, a first prompt for the task, and how we decide on day 14. 6. Risk flags: personal data, confidential information, anything customer-facing. Rules: Use plain language. I'm not technical. If AI isn't the right fix, say so. Reply in the language I answer in. Start with your first question.
2. 2

   Pick the first pilot with three questions

   The author's test: checking should take less than half the time of doing the work; start with internal work customers never see (which is why a customer chatbot is a bad first step); the owner is whoever does the task every week, not the boss.

3. 3

   Keep it or kill it on day 14

   Check two things: is the task at least a third faster, and does the owner still use it without a reminder? Both yes: keep it and start the next one.