# Email Lesson 02: A Side Project Is The Fastest Way To Learn AI Design

Source chapter:

- `docs/chapters/02-planark-side-project-ai-design.md`

## Lesson

A real side project exposes what prompt-first AI work misses: source trust, product judgment, design memory, logs, QA, and saved learning.

Planark changed how I use AI because it made the work real.

In a demo, plausible output can look impressive.

In a product, plausible is not enough.

Planark is a travel app, so a beautiful generated itinerary is not useful if the place data is weak. A confident recommendation is not useful if it ignores the actual use case.

That forced a rule:

```text
real data first -> AI after
```

This is why side projects are such a strong learning environment.

They expose the full system:

- what the product is really for
- what source material can be trusted
- where AI output becomes generic
- where the UI drifts
- where logs matter more than guesses
- what has to be saved for the next loop

The side project becomes a proof environment.

## Exercise

Pick one real project and write what AI helped with, what it could not decide, and what the workflow should save for next time.

Use this template:

```text
Project:
What AI helped with:
What AI could not decide:
Where the output was plausible but not good enough:
What source/context mattered:
What should be saved into the workflow:
```

## Artifact

Proof project note.