Every practitioner has a strong suit. Vignelli: restraint and structural clarity. Victore: raw personality and fearless expression. Bierut: intellectual rigor and whether the design thinks. Muller-Brockmann: grid as governance. Rand: concept as irreducible unit. Millman: authenticity and whether this person is real. Shaw: the room, the walk, the feeling before the content.

I extracted evaluation lenses from each of them. Not their styles. Their questions. What does Vignelli ask of a piece? Can it survive further reduction? What does Victore ask? Is this fiercely itself? What does Millman ask? Is this person real?

Those lenses became my evaluation toolkit. I run them against my own work. But the real exercise is staffing: which lens do I assign to which layer?

Here’s the practicum. Take a project. Any project. A brand, a page, a product, a presentation. Identify the structural layers: the architecture, the content, the visual system, the voice.

Now staff each layer with a practitioner. Not a style reference. A question.

The architecture might get Muller-Brockmann. Is the grid holding everything in relationship? Is the structure governing the content or is the content fighting the structure?

The content might get Bierut. Does this think? Is there an idea at every level, or is it just arranged?

The visual system might get Vignelli. Can anything be removed without losing the concept? Is the restraint producing clarity or producing emptiness?

The voice might get Victore. Is this fiercely itself? Does it sound like the person who made it, or does it sound like a template?

And the whole thing gets Shaw. Would a stranger walk into this and know who this person is before reading a single case study?

Staffing the layers is a selection exercise. You can’t run every lens on every layer. You’d be evaluating forever. The skill is reading which layer needs which question right now. A project with strong architecture but weak voice doesn’t need another grid review. It needs Victore. A project with strong personality but no structural logic doesn’t need more expression. It needs Muller-Brockmann.

I use this approach in my AI tool architecture. Each evaluation lens is a separate skill. The coordinator (the steward) decides which lenses to dispatch based on what the work needs. The dispatch logic is the staffing decision. Run the wrong lenses and you get answers to questions nobody asked. Run the right ones and each verdict tells you something actionable about a specific layer.

The exercise scales. Staff a page. Staff a brand. Staff a career. What question does each layer of your work need to answer right now? Who would you hire to ask it?

Build your evaluation panel. Not from aesthetic preferences. From the questions that matter for the layer you’re working on. The practitioners are your crate of records. The staffing decision is the DJ selecting which track this room needs next.