Course A · Capstone · Lesson 17 of 17

Capstone A: Ship Your First Thing

Put it all together — one automation or one small app, running for real, right now.

~30 min · All your Course A work nearby · Prereq: completed A3 or A4 modules

This is not a quiz. This is the moment you point everything you learned at a real problem in your business and ship something. Not a demo, not a practice run — a working tool or automation that does a useful job today. Your single win: by the end of this capstone you will have shipped one thing, it will be running, and you will know exactly what it does and what to do if it breaks. That's the whole Course A promise.

Choose your path

Path 1 — Automation (Module A3)

  • A trigger-action workflow in Zapier, Make, or similar
  • Runs on its own when the trigger fires
  • Good for: recurring tasks you do weekly or more
  • Examples: inquiry triage, meeting-notes summary, social post drafter from new content
  • Human checkpoint: review before anything reaches a customer

Path 2 — App (Module A4)

  • A one-page tool built and deployed with Lovable or Bolt
  • Accessible at a real public URL
  • Good for: a transform you do manually and often
  • Examples: client note cleaner, inquiry reply drafter, proposal outline generator
  • Human checkpoint: you review every output before using it

Pick the path that matches the problem you identified in your Course A work. If you did both modules, pick the one where you have a clear "done" test ready. You can ship the other one after — but ship one first.

The assembly checklist

Work through this before you call it done. Every item has a clear yes/no answer. If you hit a "no," fix it before moving on — or make a conscious decision that it's acceptable for v1 and write it down as a known gap.

Capstone checklist — tick every box
  1. One narrow job. State the job in one sentence with one verb. If it contains "and," scope it down. Write the sentence here before you build or deploy.
  2. Low risk if wrong. If the output is incorrect, is the worst-case outcome annoying (fixable) or catastrophic (irreversible)? If catastrophic, add a human review step before shipping.
  3. Human in the loop where it touches customers. Any output that reaches a customer — an email, a reply, a document they see — must pass through a person's eyes before it goes. No exceptions in v1.
  4. A concrete "done" test. Describe one real input and the output you'd accept. Run it now. Does it pass?
  5. You can explain how it works. Could you describe to a colleague what happens between input and output? If not, you own something you can't maintain. Spend 10 minutes with the AI translator (Lesson 15) until you can.
  6. You know how to turn it off. For an automation: you can disable the trigger in under two minutes. For an app: the URL can be unpublished or access removed. Know the off switch before it matters.
  7. It is live. Automation: the workflow is active and you've let one real trigger fire. App: the public URL loads and passes your "done" test in a fresh browser.
Done is better than perfect. A tool that does 80% of the job and is running today is worth more than a flawless spec that ships next month. Write down the 20% gap — that's your v2 backlog. Ship what works.

Definition of done

You're finished with Capstone A when all three of these are true:

Capstone A is complete when:
  1. The tool or automation is live and running (not just "works on my machine").
  2. You have run your real "done" test on the live version and it passed.
  3. You can describe in two sentences what it does and what happens if it's wrong.

What's next: Course B

You shipped something. Now the questions change. In Course B you'll take what you built from prototype to something sturdier: how to use version control so you can always roll back, how to work alongside AI coding agents on bigger features, how to read and trust the code your AI generates at production scale, and how to build the agentic toolkit — AI that decides its own steps rather than following a fixed recipe. Course A taught you to steer. Course B teaches you to build the vehicle.

Check yourself

Answer before opening — recalling it is what makes it stick.

You built an automation that auto-replies to every new customer inquiry. Is it ready to ship as-is?

No — it needs a human-in-the-loop checkpoint. Auto-replies that reach customers without a human review step are a v1 anti-pattern. Something will go wrong — a misread tone, a wrong answer, an edge case — and it will reach the customer before you catch it. Draft the reply automatically; let a person approve and send.

Your checklist item 5 ("you can explain how it works") feels awkward — do you really need it?

Yes — it's your maintenance guarantee. If you own a tool you can't explain, you can't fix it when it breaks, you can't tell a colleague how to cover for you, and you can't improve it safely. Ten minutes with the AI as translator (Lesson 15) converts "magic box" into something you actually own.

What is the one difference between what you shipped in Course A and what Course B will build toward?

Durability and scale. Course A tools are prototypes: they work, but they're fragile, lack version control, and aren't designed for production load or team collaboration. Course B adds version control, production-grade reliability, and the agentic toolkit for more complex multi-step work.

Read this next (primary source)

Building Your First AI Feature: Non-Technical Founder's Guide — Mean CEO. A complementary perspective on scoping, shipping, and keeping a human in the loop for your first AI feature — written for founders, not engineers. ~12 min. See RESOURCES.md for more.

I'm your teacher — tell me what you shipped. Describe your tool or automation in two sentences: what it does and what happens if it's wrong. I'll confirm whether your human-in-the-loop design is right for the risk level, and point you toward the Course B lesson most relevant to where you want to take it next.

Terms from this course — automation, human-in-the-loop, build brief, deploy, debugging — are all in the course glossary. Keep it as a reference as you move into Course B.