Test Drive Week + Graduation
A built dashboard isn't the same as a working dashboard.
The hardest part of M10 isn't the wiring: you finished that in 10.2 (the skeleton) and 10.3 / 10.4 / 10.5 (wiring it up). The hardest part is the seven-day stretch where the dashboard either earns its keep or quietly becomes another thing you set up and never used. Test Drive Week is the test.
Three light check-ins, spread across one normal week. Each is 5 minutes and asks the same question in three different ways: did the dashboard actually help, and if not, what specifically broke? No journaling. No reflection essays. Three task-based status pings.
What the test drive is: and isn't
It is: a structured 7-day rhythm to confirm your three scheduled tasks fired, your library is reachable, and the Coach is producing useful output. It isn't: a feelings check, a habit-tracker, or a "how do you feel about your relationship with AI" exercise. We're checking the system, not you. If something doesn't fire, that's a wiring fix, not a personal failure.
Pick a normal week to start
Don't run this during finals week, application crunch, or the week you're traveling. The point is to test the dashboard against your typical schedule, not a peak one. Most students start on a Sunday (so Day 1 lines up with a Sunday Reset firing) and run through the following Saturday.
Three check-ins, three distinct moments
Day 1 (Sunday or Monday): did the system start firing? Day 4 (mid-week): what's working, what's friction? Day 7 (end of week): what survives the semester, what gets edited?
The Day 1 plumbing check
- Open Claude. Look for the Sunday Reset output (or your earliest-firing task) in the message list. Did it land? Yes → continue. No → straight to the capstone-troubleshooting bonus page for the "scheduled task didn't fire" recipe.
- Open the output once. Skim it. Did it produce something usable, or did it look obviously wrong (referenced last term's classes, missing your real calendar, generic "here's a planning template")?
- Open your Coach Cowork Project. Confirm your starter prompts are reachable from the chat input: can you see at least 3 of your 10 saved prompts in a couple of clicks?
- If anything in steps 1–3 didn't work, note exactly which one in the workbook on Screen 3. The fix recipes are in the troubleshooting bonus page.
- If everything worked, write the date in the workbook. You're cleared to keep going.
Day 4 + Day 7 · the mid-week and the close.
Day 4 maps each piece to one of three states (working / friction / unused) and you pick one highest-leverage tweak to make. Day 7 closes the test drive: keep / edit / cut for each piece, with at most 2 edits across the whole dashboard.
The Day 4 working / friction / unused triage
- List your dashboard pieces from Panels 2–4: the Coach, the three slots, the 10 saved prompts. (Group prompts by Cowork Project; you don't have to list all 10 individually.)
- For each piece, mark one of three: ✅ Working (fired on time, useful, you used it), ⚠️ Friction (output needs editing, prompt needs tightening, or you keep ignoring it), ❌ Unused (didn't fire, didn't get opened, or you forgot it existed).
- Pick the one piece in "friction" or "unused" with the highest payoff if fixed. Note it. That's your one tweak before Day 7.
- Don't fix more than one thing. Three days between check-ins is to let patterns emerge: too many simultaneous tweaks make it impossible to tell what helped.
The Day 7 keep / edit / cut review
- For each dashboard piece, decide: KEEP as-is (running cleanly, leave it alone), EDIT (tighten one specific thing: prompt wording, schedule time, standing brief), CUT (pause this piece: slot disconnect or saved-prompt remove; M9.1's "ignored 3 times in a row" rule applies, just compressed to one week).
- Make at most 2 edits. If you find yourself making 6 edits, you don't have a tuned dashboard yet: restart the test drive next week with the 2 highest-leverage edits and re-run.
- Fill in the workbook 7-day log on the next screen. Last big workbook panel; piece 5 of your keepsake.
- Re-run anything you cut after 4 weeks if you change your mind. Cuts aren't permanent. The dashboard is yours.
"Friction" is more common than "unused"
Most pieces of a fresh dashboard land in friction, not unused. Friction is the input you need to tighten the prompt body, the standing brief, or the schedule. It's normal week-1 information, not "the dashboard isn't working." If a piece is fully unused on Day 4, the question is: was it solving a real pain, or a pain I imagined I had?
Workbook · your 7-Day Log + what passing looks like.
The honest record of your test drive week. The keepsake's last piece. Below it, the realistic verdicts students actually hit: so you can compare yours and see if the system held up.
Your 7-Day Log
Test-drive results, keep/edit/cut decisions, your one-line verdict.
All five panels filled in (1 from 10.1, 2 from 10.3, 3 from 10.4, 4 from 10.5, 5 from this lesson):that's the "My AI Life System" record from the Capstone spec, completed.
Realistic Day 7 verdicts
- "Sunday Reset is great. Slot 2 needs tightening. Slot 3 I never touched.": Sunday Reset is the durable one; the rest takes more iteration. Action: edit Slot 2's prompt body, decide whether to keep or cut Slot 3.
- "Sunday Reset is great. Both other slots are working. Library felt clunky.": Action: trim 2–3 saved prompts; the library only needs to hold the ones you actually run.
- "Everything's running but the output feels generic.": usually a Coach issue, not a slot issue. Action: revisit the standing brief from 10.3 and add 1–2 sentences of personal specificity.
- "Slot 1 didn't fire reliably.": almost always a connector or timing issue. Action: hit the troubleshooting bonus page recipes (calendar off, AM/PM mistake, account on free not Pro).
- "I forgot to use it for two days mid-week.": common and fine. The dashboard is supposed to be there when you remember; it's not supposed to occupy headspace when you don't. Action: nothing. The system held up.
You built a working system. That's what we set out to do.
Take this in for a second: you have a tuned Coach Cowork Project, three Cowork Projects total in your dashboard, three scheduled tasks (one of them the M9 flagship), ten saved starter prompts, and a five-panel record of how it all came together. You ran it for seven days, kept what worked, edited what didn't, cut what failed. That's an actual system running on your account, doing real jobs in your real week.
What you have
A Claude Pro account with: three Projects (Coach + Applications + Personal), three scheduled tasks (Sunday Reset locked in, one Pre-Exam Watch tuned to your highest-friction class, one of the variants or empty), ten saved starter prompts indexed in your workbook, two connectors on (calendar + Gmail), five workbook panels documenting the whole build.
The calendar reminders that make the dashboard a year-round tool.
Students usually finish a course and never come back. With this course, you built a system that gets better and more refined the more you use it. Certain points in the academic year (and beyond) will be when you need it most. Set the reminders below in your calendar right now, and the dashboard will surface itself when you actually need it instead of slowly fading into your sidebar.
Pull-this-out-before · academic and career events to set as recurring calendar reminders
- 2 weeks before each midterm season: pull out 10.4 (re-confirm Pre-Exam Watch is tagged correctly for this term's exams) and the troubleshooting bonus (re-run the stale-Coach refresh).
- 1 week before finals: pull out 10.4 (Slot 2 Pre-Exam Watch on every final this term) and the M3.5 study-stack prompt out of your library.
- Start of each new term: pull out 10.3 (re-tune the Coach with new syllabi, replace last term's PDFs) and M7.4's end-of-semester archive dance.
- August 1st (or your application year's equivalent): pull out the Applications Cowork Project from M6.2 and run it through 10.3-style tuning. Set Slot 3 to App Deadline Tracker if it's empty.
- 2 weeks before a major application deadline: pull out M6.5 (resume / cover / LinkedIn workflow) and M6.3 (personal statement) from your library. Triage what's due.
- 1 week before a real interview: pull out M6.6 mock-interview prompt from your Applications Cowork Project. Run a few rounds in voice mode the day before.
- 2 weeks before study abroad / a big trip / moving: pull out the dorm-roommate and travel-prompts bonus pages.
- End of every semester: pull out M7.4's archive dance + M9.1's connector audit + the every-semester re-audit habit. Trim what's not earning its keep.
Set 3 right now, not 8
Pick the three from the list above that are most likely to come up in the next 6 months for you. Add them to your calendar as recurring reminders. Don't try to set all eight. Calendars full of dashboard-pull reminders become noise; three is enough to keep the system in your peripheral vision when you'll need it. The other five you can add later when they're actually relevant.
The format for the reminder itself
Title each reminder: "M10 dashboard · pull this out before [event]." Description field: a one-line note about which lesson(s) to revisit and which prompt(s) to run. That way when the reminder fires, you don't have to remember why you set it: past-you wrote that down.
The close.
Ten modules. Fifty-something lessons before this one, plus the six in M10 you just finished. Most students will not have read every word of all of them, and that's fine: the course was designed to be skimmed back to. The dashboard you built is what makes the rest of the course retrievable: when you forget what M3.4 said about Explain-It-Back or what M6.3's 4-phase pattern was, you'll come back, find the lesson, run the prompt, and the muscle returns. That's the design.
Module 10 · Capstone: Your Student Life AI Dashboard (recap)
- 10.1:Pick Your Pain Points. The dashboard inventory, top 3 pains, the pain-to-module crosswalk. Signature Decisions + Map panel.
- 10.2:Build Your Dashboard Skeleton. The Cowork Project, the Blueprint interview, the one-prompt artifact build, the cross-Project awareness rule. The dashboard itself: live, in your sidebar.
- 10.3:Upgrade Your Coach. Re-tune the four staple files + the Capstone-edition standing brief + the one-time tuning move. Coach Tuning panel.
- 10.4:Three Scheduled Tasks. Slot 1 Sunday Reset (locked) + Slot 2 Pre-Exam Watch (highest-friction class) + Slot 3 (your pick or empty). Three Scheduled Tasks panel.
- 10.5:"Start Here" Prompt Library. 10 saved starter prompts split between Coach + Personal Cowork Project. 10-Prompt Library panel.
- 10.6:Test Drive + Graduation (this lesson). Day 1 plumbing / Day 4 triage / Day 7 keep-edit-cut + the close. 7-Day Log panel.
One last time · the Honest Work Code, applied to your dashboard at the course level
Rule 1 · You stay the author of your own work and life. The Coach has standing order 4 hard-coded against writing for you. The scheduled tasks are typing-tax cuts, not deciding-tax cuts: Sunday Reset proposes a Week Map; you edit it before Monday. The library is named in your shorthand because you chose those ten. The Voice Profile is so the Coach can match your voice when editing yours, not so you can pass off Coach output as your voice. Every flagship beat from earlier modules: the Hallway Test from 4.1, the Dinner Table Test from 6.1, the Citation Trap line from 5.1, the "default off; on must earn its slot" from 9.1: holds in this dashboard, end-to-end.
Rule 2 · Your work survives scrutiny. You documented this build in five workbook panels. You named your scheduled tasks. You connected calendar and Gmail (and only those two). You followed M4.5's process trail when writing, M5.2's source-checking when researching, M6.7's helper-doc rule when asking for recommendation letters. If a parent / professor / IT admin / future-you ever asked "what does AI do for you in school?":you have a one-paragraph answer and a workbook to show. That's the bar.
Rule 3 · Respect the rules of every room you walk into. None of your scheduled tasks fire during exams. You've checked your school's policy (M4.2) and read each professor's syllabus rules. You haven't run AI live during a real interview (M6.6). You haven't drafted your own recommendation letter even when a prof asked for "something to start from" (M6.7). The dashboard supports preparation, not performance-time crutches. That distinction is the entire architecture.
The dashboard is yours, not ours
Edit anything. Cut what's not earning its keep. Re-tune the Coach as your tone changes. Add a Slot 4 if you find yourself needing it. Drop Slot 3 if it's been empty for two months. The shape of the dashboard you graduate with is not the shape it'll be in a year: and it shouldn't be. The system is there to serve your life, not the other way around.
Show your dashboard if it'd help a friend
If a friend asks you "what does using AI for school actually look like for you," walk them through your dashboard. Open your Coach Project, show them the staple files and the standing brief, point at your scheduled tasks running, scroll through your saved starter prompts. It's the most honest answer to that question, way better than abstract advice. If they want to build their own, the course is here for them too. You can also send them this sample of one student's dashboard — it gives them the picture in 30 seconds.
Cowork can do more than this course covered.
The student-specific workflows are what we taught. The bonus page What Else Cowork Can Do is a menu of capabilities that didn't fit in the main lessons: organizing files, batch-drafting replies, cross-app workflows, and more. Bookmark it for when you have a new task to throw at it.
You finished. The dashboard is built.
The workbook panels are filled; the calendar reminders are set; the Coach is tuned; the three slots are wired; the library is reachable. Come back to the curriculum whenever you need a specific lesson: the dashboard is the index, the bonuses are the reference library, and the lessons stay where you left them.
Back to the curriculum →