Week 7: Practice Project + 5 ways I use AI to improve my dbt skills

Practice Project Checkpoint 2 is live with a wealth of tasks to improve your understanding in how to set up models & sources, use node selectors and check artifacts.

Part of the “Mastering dbt” series. Access to the full Study Guide. Let’s connect on LinkedIn!

We’ve reached the end of Checkpoint 2 with a very “meaty” Practice Project. In this stage, we built out the intermediate and marts layers, enhanced the configuration of sources and models, and familiarised ourselves with node selectors and artifacts.

One highlight for me was creating a business case for our data. Framing the models around a story really helps to think logically and design them with purpose.

That said, here’s a small rant: I found that applying the official documentation to the project wasn’t always straightforward. Some configurations turned out to be trickier than the docs suggested, and it wasn’t always clear what could be declared where.

I ended up relying on AI and some extra research to clarify a few points — which, on the bright side, gave me the idea to share how I use AI as a learning tool.

Next week, we are covering one of my favourite dbt topics: refactoring SQL! I love transforming that looong SQL script into carefully organised and neat models.

---

How I use AI to support my learning

AI can be your friend in your learning journey, but you have to use it wisely. Use it too much and you’ll end up dependent on it and possibly misinformed. Don’t use it at all, and you might miss opportunities to enhance your understanding.

Here are 5 ways in which I use AI while navigating the dbt documentation and working through the Practice Project:

1. Seeing real examples
   When the documentation is vague, I ask AI to show examples of dbt features in action. Seeing a config or macro applied in context makes it much easier to understand.

2. Clarifying confusing docs
   If something in the documentation doesn’t make sense, I ask AI to explain it. I can even share the documentation link and point out exactly what’s unclear.

3. Answering questions not covered in the docs
   AI is great for linking and comparing different features. For example, it helped me understand the differences between sql_header, query-comment, and hooks in dbt.

4. Highlighting best practices
   Documentation often shows multiple ways to configure something, but it doesn’t always say which approach is recommended. I ask AI for advice based on my use case and best practices.

5. Debugging errors
   “What the hell does this error mean?” — a classic. AI can help interpret cryptic error messages and suggest what to check next.

Extra tip: I keep all my Practice Project-related messages in a dedicated chat. That way, AI has context about the data I’m working with and can give more relevant answers.

A word of caution: AI is not infallible. Use it to support your learning — alongside the documentation, experimentation, and your own judgment — rather than as a replacement for them.

---

Have the study notes been useful to you so far? How are you getting on with your self-paced learning journey? I would love to hear from you here or on LinkedIn!