Your 30-Day Plan

What to do after the course finishes, so the skills survive contact with a normal working month.

The failure mode is predictable: the cohort ends, the momentum goes, and in six weeks you are back to using AI as a search engine. This plan is deliberately small enough to survive a bad month.

The commitment

Fifteen minutes a day, four deliberate builds, one review a week. That is all.

Not "use AI more". A specific task, on a specific day.

Week 1 — make it a habit

The goal is reflex, not capability. You already have the capability.

Day	Do this	Minutes
1	Put your Digital Twin into your assistant's custom instructions	15
2	Use AI on the first thing today you would normally do by hand	15
3	Take a prompt that half-worked and improve it. Save the improved one.	15
4	Set up your daily briefing, or improve the one you have	20
5	Ask it to critique something you wrote. Do not let it rewrite.	15
6	Deliberately try to make it get something wrong. Note the failure mode.	10
7	Review: which of these will I keep doing?	10

Week 2 — build one thing

One automation. Small. See the capstone selection worksheet for choosing.

Day	Do this
8	Pick it. Write the scope, including what it will NOT do.
9	Record the baseline — time it by hand, once.
10	Build the trigger and filter only. Test.
11	Add the AI step. Test ten times.
12	Add the output and the error path.
13	Run it live. Watch it.
14	Review: what broke, and what did I learn?

Week 3 — make it durable

The week most people skip, and the reason most builds do not survive.

Day	Do this
15	Write the one-page SOP
16	Organise your prompt library. Delete what you never use.
17	Run the privacy checklist across every tool
18	Measure: what did week 2's build actually save?
19	Fix the thing that annoyed you most about it
20	Show it to someone. Watch where they get confused.
21	Review: is this still worth running?

Week 4 — teach and extend

Day	Do this
22	Teach one technique to one person. Teaching exposes what you do not understand.
23	Try one tool you have not used
24	Revisit your tool stack — anything to drop?
25	Pick the next thing to build. Write the scope.
26	Update your Digital Twin with what has changed
27	Write up build one as a short case study
28	Review the month

Days 29–30: the honest assessment

Time I actually spent: [HOURS]
Time I actually saved: [HOURS PER WEEK]
What I built:
What I still do by hand that I should not:
What I automated that I should not have:
The habit that stuck:
The habit that did not, and why:
One thing for next month:

That fifth line matters. Some things are worse automated, and noticing which is a real skill.

After the 30 days

The weekly fifteen minutes. One slot, every week, permanently:

• Anything I did three times this week that could be a prompt or a workflow?
• Any automation misbehaving?
• Any prompt that keeps needing the same correction? Fix it at the source.

The quarterly hour.

• Re-test your automations. Models change under the same name.
• Review your tool stack and costs.
• Update your Digital Twin.
• Turn off anything unused.

Staying current without drowning

The information volume is the problem, not the scarcity.

• Two sources, not ten. Pick two you will actually read.
• Ignore benchmarks. They move weekly and rarely change what you should do.
• Try a tool only when you have a task for it. Tool-first exploration wastes time.
• Follow capabilities, not releases. "Models can now do X reliably" matters; version numbers do not.
• Re-read your own notes quarterly. Half of what you wrote down you have already forgotten.

The thirty-day tracker

Day	Done	Minutes	Notes
1
2
…

Print it or keep it open. The point of the tracker is not discipline; it is that at day 30 you have evidence rather than an impression.

The one thing

If you do nothing else on this page: put a recurring fifteen-minute slot in your calendar, weekly, called "AI review".

Everything else on this page is optional. That slot is what makes the difference between having done a course and having changed how you work.