Lesson 5 of 5 · 6 min · final lesson

Put it to work.

You've got the picture now: what Claude Code is, how to set it up safely, the jobs it does well, and how to keep people in the loop. The last step is the one that actually changes anything, because reading about an agent teaches you nothing until you've watched one work on your own code. This week, you trial it. Here's the calm way to start.

Pick a few well-scoped jobs to trial

Don't reach for a sprawling rewrite to prove the point. Pick a handful of small, bounded jobs with clear done criteria, the kind from lesson three: a checkable result, low blast radius, and no production secrets in sight. Good first candidates for almost any codebase:

Add test coverage to a module that's light on it.
A tidy refactor with the suite as your safety net: rename, extract, split a fat file.
A small bug with clear reproduction steps, so it can write the failing test and fix it.
A codebase question to warm up: ask it to explain a module or trace where something is handled.
A small internal tool or one-off script that saves the team a recurring chore.

Write your shortlist down before you start, and make sure your repo is ready: a CLAUDE.md with how to build, test and lint, your permissions set so you approve as it goes, and branch protection with a green CI as the backstop. A few real tasks beats a vague intention to "try the agent" every time.

Run the loop: scope, run, review, merge

For each job, the rhythm is the same, and it's quick:

Scope. Write a tight brief: one clear goal, a definition of done, and the boundaries. If you can't scope it cleanly, break it down first or keep it for yourself.
Run. Hand it to the agent and let it work, approving actions as they come up. Use the cloud option or run a couple in parallel if the jobs are independent, and lean on subagents for a big job: the agent can spin up focused helpers to explore or handle sub-tasks, then bring the results back together.
Review. Read the diff like any contributor's. Is the approach sound? Anything off? Green tests are a start, not a sign-off. This is where your judgement goes in.
Merge. If it's good, merge it through your normal protected flow. If it came back off, tighten the brief or improve the project instructions and run it again. That feedback is how both of you get better.

That last habit is the compounding bit. Every time you sharpen the brief or add a missing note to your CLAUDE.md, the next task lands closer to right. Over a few weeks the agent quietly clears a real slice of the well-shaped work, and your people spend more of their hours on the parts that need them.

The mindset to keep

The whole thing rests on one balance. Let the agent carry the bounded, checkable jobs, and keep your hands on design, on the risky paths, and on the decision about what's worth building and what's worth shipping. Your review is still the thing that keeps the work good and safe, so it never gets skipped no matter how fast the diffs come back. Get that split right and the speed is real while the risk stays low.

Keep the summary handy

To make this stick, grab the free one-page summary: what Claude Code is, the scoping rules, the permissions and secrets line, the MCP note, the review and CI guardrails, and a space for your first tasks. Print it, stick it by the desk, and you've got the whole course in arm's reach.

Your week, in one line: pick a few small, well-scoped jobs, run the scope, run, review, merge loop on each, and tighten the brief when something comes back off. Keep a human reviewing every change and the secrets well clear. That's how Claude Code goes from a thing you've read about to a quiet, useful part of how your team ships.

Quick check

A few quick questions to lock it in. No marks recorded, just for you.

Q1.Before you let it write, how should you get a feel for a first task?

Give it auto-approval and a sprawling rewrite to prove the pointUse plan mode on a small, well-scoped job so it lays out an approach before touching any filePick whatever touches production secretsChoose whatever looks most impressive in a demo

Plan mode lets you approve the approach risk-free, and a bounded, checkable job is the safe way to learn the workflow.

Q2.What's the loop to run for each task?

Hand it over and merge whatever comes backScope it, let it run while you approve actions, review the diff, then merge if it's goodSkip review and rely on the agentStart from scratch every time with no CLAUDE.md

Scope, run, review, merge. Tighten the brief or the CLAUDE.md when something comes back off, and keep what works.

Q3.What's the mindset to keep?

Loosen permissions to auto-accept and step awayLet it carry the well-defined work while you keep your hands on design, review and the risky pathsAvoid agents until they're flawlessDo it all by hand to be safe

Even with the permission dial wide open on safe work, your judgement on what to build and what to ship is the advantage.

Answer all the questions to continue.

Course complete

That's the course done. Nice work.

You've got the full picture now: what Claude Code is, how to set it up safely, the jobs it does well, and how to keep people in the loop. Here's what to do next.

Answer the quick check above to unlock this.

Free download

Now put it to work

Turn all this into action with the fill-in workbook: spot where a coding agent helps, scope your first tasks, set the permissions and guardrails, and write your review and secrets policy. Pop your email in and work through it with your team.

Company

Your name

Also send me the occasional practical AI tip for business. (optional, unsubscribe any time)

Your email gets you the workbook, that's it. No spam, and I never share your details.

Prefer the quick version? Read the free one-page summary →

You're in.

Your fill-in workbook is ready. Open it below, then type straight into it or print it to work through with your team.

Open your workbook

Pick up anywhere

Save your progress

Pop your email in and we'll send you a link to pick up where you left off, on any device. No account needed.

Company

Also send me the occasional practical AI tip for business. (optional, unsubscribe any time)

Just for the link to your progress. No spam, and I never share your details.

Saved.

Check your inbox for a link to continue on any device.

Lesson 4 Keeping humans in the loop Finish Explore more courses

Keep going

Keep learning at your own pace.

The other free JDCS courses go broader on AI and automation, and the guides dig into the parts that matter most to how you build.

Browse the courses

Connect with me

Want a hand putting an agent to work in your team?

Tell me a little about how your team ships and I'll come back with a plain-English read on where a coding agent would help most, and the guardrails to put around it first. The first conversation is free, and you keep the plan either way.

Step 1 A few quick questions

Company

Where's your time actually going?

Pick whatever rings true: the stuff you'd happily never do again.

Quoting & proposalsAdmin & data entryCustomer enquiries & follow-upsScheduling & bookingsReporting & numbersSomething else

What are you running the business on?

The tools you already use. It tells me what we'd be plugging into.

Xero / MYOB / QuickBooksServiceM8 / TradifyA CRM (HubSpot, etc.)Google WorkspaceMicrosoft 365Shopify / WooCommerceStripe / paymentsMostly spreadsheets & emailSomething else

Are you using AI for any of it yet?

No wrong answer. It just tells me where you're starting from.

Yes, by copy-pasteA littleNot yetNot sure where to start

Tell me about the business.

What do you do?

How many of you?

Just me2–1011–5050+

Where are you up to?

Just exploringLooking to start soonGot a specific project in mind

If this worked, what would it free you up for? (optional)

Where do I send the plan?

I'll come back with a plain-English read on what's worth automating. No obligation.

Your name

Phone (optional)

Anything else? (optional)

Got it, thanks.

That's landed with me. I'll read it properly and come back with a plain-English plan, usually within a few hours and always within one business day.

Success. You've finished the course.

Nice work. You know what Claude Code is, the jobs it does well, and how to set it up safely with a human on the things that matter. Want to take it further?

Keep learning