Mood and tone before the first screen

Why it starts by hand

Mood and tone is the feel of a product: calm or loud, playful or serious, premium or friendly. It is the first thing a client reacts to, and the hardest thing to change later.

That is why I make it by hand. When the client asks "why does it look like this?", I need a real answer: something I chose, for a reason. "The AI suggested it" is not a reason.

Start from generated screens and you tend to land on the average of everything. Fine, forgettable, and nothing I can defend.

A few directions, not ten

I make a couple of directions that are clearly different, say clean and minimal versus bold and colourful. Each one gets a few real pieces: colours, type, one key screen. Enough to feel it, not a full design.

Then we pick together. Talking through why one feels right is half the value. It tells me what the client actually cares about, which the brief rarely says out loud.

Lock it, then make it a system

Once the client picks, the direction becomes variables: colours, type, spacing, radius. That is what makes it last past the first screen, and what lets a team keep it consistent. (The token part is note 14.)

Then let AI extend it

Now AI is useful. I show it the chosen direction, the tokens and the key screen, so it learns the style from my work. Then it helps with the system work: palettes, more screens in the same style, naming. We improve it together, and I check each piece against what the client signed off.