Every tool we use now can produce something finished-looking in minutes. A logo generator spits out six options before your coffee cools. An AI image model can render a full brand concept, moodboard included, from a two-line prompt. Even inside Figma, templates and component libraries mean a competent designer can build something that reads as a comp almost as fast as they can think of it.
That speed is not the problem. The problem is what it quietly replaces: the slow, ugly, deliberately unfinished phase where the real thinking happens.
Resolution Isn’t the Same as Rightness
A finished-looking draft is persuasive in a way a sketch never is. Smooth gradients, real typography, a mocked-up browser chrome around a screen — all of it signals decided, even when nothing has actually been decided yet. That’s the trap. Fidelity and correctness are different axes entirely, and the tools that have gotten faster are almost all tools for fidelity, not for correctness.
We still start concepts with pencil or marker on paper, or at best a grayscale wireframe with boxes and arrows, before anyone opens a screen to build anything real. Not out of nostalgia. Because a rough sketch is cheap to throw away, and a polished comp isn’t — psychologically, even if the file itself costs nothing to delete. Once something looks finished, killing it feels like a loss. Killing a thumbnail sketch feels like Tuesday.
This is the structural gap between us and the faster alternatives. A freelancer juggling five clients has every incentive to show up with something that already looks like progress — a comp, not a question. An AI tool has no incentive at all; it was built to produce a finished-looking artifact on the first try, every time, because “finished-looking” is the only thing it knows how to grade itself on. Neither is set up to protect the phase where you’re allowed to be wrong in public, in front of the client, cheaply.
What a Sketch Phase Actually Catches
The point of low-fidelity work isn’t the drawing. It’s what drawing badly, quickly, and a dozen times over forces into the open before anyone’s attached to an answer:
- Wrong problem, right-looking solution. A beautiful logo for a positioning nobody’s agreed on yet is just an expensive distraction. Sketching keeps the conversation on the idea, because there’s nothing polished enough to argue about instead.
- Hierarchy problems a comp will hide. A rough layout with boxes and labels exposes whether information actually flows in the right order. Add real type, color, and imagery too early, and the polish papers over a structure that doesn’t work — you won’t notice until it’s live.
- The second-best idea, killed too early. Teams that go straight to one polished direction rarely generate more than one or two real alternatives — finishing anything is expensive, so you stop early. Teams that sketch can cheaply produce ten directions and genuinely compare them, which is the only way to know your favorite is actually the best one and not just the first one.
None of this shows up in a client-facing deck. It’s invisible by design. But it’s also exactly the part of the process that a faster, shinier-looking alternative has no good way to compress, because compressing it is the whole shortcut.
The Business Case for Slow
This isn’t just craft romanticism. McKinsey’s Business Value of Design research, which tracked more than 100,000 design actions across 300 public companies over five years, found that the top-quartile companies on its Design Index — the ones with the most rigorous end-to-end design processes, not just the best-looking final output — grew revenue 32 percentage points faster and total shareholder returns 56 percentage points faster than their industry peers. The differentiator in that research wasn’t visual polish. It was process discipline: testing early, testing often, and across the full product or brand experience rather than just the final layer anyone outside the room ever sees.
That tracks with what we see client-side. The brand systems that hold up three years later, through a product launch, a leadership change, an acquisition, are almost never the ones that looked most impressive in the first client presentation. They’re the ones where somebody was willing to kill four directions before the fifth one survived contact with a real business problem.
Slow Down Where It’s Cheap
The paradox is that sketching feels like it’s adding time, when it’s actually the only phase where speed is free. A tenth thumbnail costs you two minutes. A tenth full comp costs you a day, and probably a client who’s now emotionally attached to the ninth one. If you want a faster, better brand process — and everyone does — the fix isn’t skipping the sketch. It’s doing more of it, earlier, before anything on the screen looks finished enough to stop questioning it.
Novelty is the easy part to buy today. Discipline about when to stay rough is the part that’s actually in short supply — and it’s the part no prompt box has figured out how to fake.
