A classic web project passes through five pairs of hands: strategist, art director, designer, front-end builder, developer. At each handover, a little of the intent evaporates. The brief gets diluted, the mockup interprets the concept, the build interprets the mockup. The result resembles the original idea from a distance. Rarely up close.

The cost of seams

Every seam between two trades is a small leak. You lose meaning, time, precision. You re-explain, you readjust, you renegotiate. The client pays for that friction without ever seeing it on the quote. And yet it is there, in the back and forth, in the compromises, in the final gap between what was wanted and what was delivered. Nobody is at fault, by the way: that is what the division of labour does.

One pair of hands

Since 2004 I have learned every link because the trade pushed me into it, from Flash to AI. Today I hold the chain from one end to the other: the angle, the art direction, the design system, the build, the deployment. Not out of a taste for controlling everything. Simply, an intent survives better when it does not change carrier.

The flow of a five-handed project, then that of a chain held from one end to the other Two stacked rows. On top, five aligned boxes named strategist, art director, designer, front-end, developer, each linked to the next by an arrow. Below them, a horizontal band starts thick on the left and loses height at every interval between two boxes, marked by a dashed vertical stroke; by the time it reaches the right it is markedly thinner. At the bottom, a second band of the same length keeps the starting height of the first one from end to end, with no interruption at all. A CLASSIC WEB PROJECT FIVE PAIRS OF HANDS strategist art director designer front-end developer a little of the intent evaporates ONE PAIR OF HANDS NO SEAMS from one end to the other
The route does not get shorter, it is the handovers that disappear.

What that changes

A decision taken upstream stays legible all the way into the CSS. When I choose a type scale, I already know how it will live in clamp() on mobile. When I set a concept, I know what it will cost to produce, which keeps me from promising what I will not deliver. Strategy and execution are not two worlds: it is the same line, held without a break.

The same page redrawn at each handover On top, five page silhouettes at the same scale, aligned from left to right and linked by arrows. Each holds the same three blocks: a banner at the top, a large block on the left, a smaller block on the right. From one silhouette to the next, the banner shortens and drops, the large block shrinks, the small block grows. At the bottom, an enlarged silhouette overlays the two extreme states: the blocks of the first one as dashed outlines, those of the last one as accent fills, with three thin strokes joining the corners that correspond. THE ORIGINAL IDEA THE RESULT at each handover FROM A DISTANCE RARELY UP CLOSE the original idea the result
None of these steps is at fault, it is their number that opens the gap.

AI makes that possible at the scale of one person. It takes the execution, I keep the continuity. One project, one intent, no seams.