# Pilot pack — getting three recruiters to test this course *The one check that automated verification cannot replace.* Everything in this course has been checked for broken links, accurate dates, readable language and working spreadsheets. None of that tells you whether Lesson 4.4 actually works when someone runs it on a Tuesday afternoon with a real requisition open. Three readers will find things no amount of checking can. This pack is what to send them. --- ## Who to ask Three people, deliberately different: | Reader | Why them | |---|---| | A recruiter with 2–5 years' experience, no AI tooling | The core audience. If it does not work for them, it does not work | | Someone who hires but is not a recruiter — a founder or hiring manager | Finds the assumed knowledge nobody else notices | | A recruiter who already uses an AI screening tool | Finds where the course is wrong or out of date, rather than unclear | Pay them or trade something. Free favours get done badly and late, and you need this done properly once rather than half-done three times. --- ## What to send them > I've written a free course on using AI in recruitment — 12 modules, 73 > lessons, about four hours of reading plus hands-on exercises. > > Before I promote it, I want to know whether it actually works for someone > who isn't me. Would you spend about two hours on it and tell me where it > breaks? > > I'm specifically not looking for encouragement. I want to know where you got > stuck, bored, or lost — that's the useful part. > > [link] --- ## What to ask them to do Not "read the whole thing". Give them a specific route: 1. **Read Module 1 completely.** Five lessons, about twenty minutes. This is where people decide whether to continue. 2. **Do Lab 3** — screening the 500-CV dataset, in Module 4, Lesson 7. This is the anchor exercise and the one most likely to break. 3. **Read one module that matters to their job.** Their choice. 4. **Skim the glossary and the templates.** That is about two hours. Asking for more gets you nothing back. --- ## The six questions Send these, and ask for short answers rather than a report. 1. **Where did you stop reading, or want to?** Name the lesson. This is the most valuable answer in the set. 2. **What did I assume you already knew?** Any term, tool or idea that was used before it was explained. 3. **Did Lab 3 work?** How long did it actually take, and where did you get stuck? 4. **What did you already know that I explained anyway?** Tells me where the course is condescending or slow. 5. **What would you have wanted that isn't there?** 6. **Would you send this to a colleague? If not, why not?** --- ## Two questions not to ask **"Did you like it?"** You will get yes. It tells you nothing and it makes the next question harder to answer honestly. **"Any feedback?"** Too open. People default to praise plus one small typo. --- ## What to do with the answers | If two or three readers say | Then | |---|---| | They stopped at the same lesson | That lesson is the problem, not them. Rewrite it | | The same term was unexplained | Add it to the glossary and define it at first use | | Lab 3 took much longer than stated | Change the stated time. Do not defend it | | A module felt irrelevant to them | Check the audience paths point them elsewhere | | They would not share it | Ask what would change that. It is the honest measure of quality | **Change the course, not the reader.** The instinct on receiving "I got lost in Module 5" is to explain why Module 5 is necessary. That instinct is the reason most courses stay hard to read. --- ## Track it | Reader | Sent | Returned | Stopped at | Top issue | Fixed? | |---|---|---|---|---|---| | | | | | | | | | | | | | | | | | | | | | Log what you changed in `changelog.html` so the readers can see their feedback landed. People who see that give you feedback again. --- ## When to do it Before promoting the course anywhere. A course with a known rough edge in Module 4 is fixable in an afternoon. The same rough edge found by two thousand readers is a reputation.