Look at you, going the distance.
You made it to Module 10. That means you've covered everything from our AI philosophy to automation workflows to department-specific playbooks. Now the final step: make something real.
Your final project
Choose one of the options below. Do the work. Then share it in #ai-and-automations on Slack.
This isn't a formal submission or a graded assignment. It's a chance to use the full set of skills you've built in one real deliverable, and to show the rest of the team what good AI-assisted work looks like.
Choose your deliverable
What to do: Either create a new Claude Project for a client who doesn't have one, or audit and significantly improve an existing one.
Requirements:
- Project instructions that are specific, not vague (no "write in a professional tone")
- At least three knowledge files uploaded as .TXT
- Documentation of what you changed or added and why
- Three test prompts with the output compared against approved work
Share in Slack: A screenshot of the project instructions and a brief note on what changed and what it should improve.
What to do: Write a real content piece for a client using the full OO five-step writing workflow. Blog post, email, social captions (a full week's worth), or a client proposal section.
Requirements:
- Use a CRISP prompt and document it
- Run the banned words check and note what you replaced
- Rewrite at least 30% by hand
- Confirm it passes the "read it out loud" test
- If it's for a real client, it should be ready to send to the client for review
Share in Slack: The final piece plus a quick note on your process. What did you change from the first draft and why?
What to do: Identify a real repetitive task at OO, design the automation that would handle it, and write it up clearly enough that someone could build it.
Requirements:
- Clear problem statement: what task, who does it, how often, what's painful about it
- Trigger and action defined clearly
- Apps involved listed
- Estimated time saved per week
- Any approval steps or human-in-the-loop moments identified
Share in Slack: The full proposal. Format doesn't matter as long as it's clear. A Notion doc, a Google Doc, or even a well-structured Slack message works.
What to do: Build a reusable library of 8 to 12 prompts for your department or a specific client using the CRISP framework. These should be prompts that you or your teammates would actually use.
Requirements:
- Each prompt follows the CRISP structure
- Each prompt includes the expected output format
- At least one prompt per content type you commonly produce
- Prompts tested against a real Claude Project and outputs reviewed
Share in Slack: The Google Doc or Notion page with your library. Tag Eliza so she can add it to the official prompt library.
What to include when you share in Slack
When you post your final project in #ai-and-automations, include:
- What option you chose and what you made
- Which modules informed the work
- One thing that worked better than you expected
- One thing you'd do differently next time
These reflections are genuinely useful for the team. What you learned the hard way is probably useful to someone else.
What happens after this
Finishing this course is a T2 KPI, which means it counts. But more than that, it means you now have a foundation most of your peers don't have. You know why we use AI, how to use it well, how to keep it from making your work sound robotic, and how to push it further when you're ready.
AI moves fast. This course will need updates. If you find something that's wrong, outdated, or missing, post it in #ai-and-automations. Help us keep it accurate.
And if you end up building something interesting using what you learned here? Tell us about that too.
Good work, Optimist.
Now go make something. 🎉
Mark this module complete after you've shared your final project in #ai-and-automations.