Module 2 · Your Logo, Part 1: Find "The One"

Make it better than you can see

Ask AI “is my logo good?” and it says yes. Every time. Chatbots were found to affirm you roughly 49% more often than a person would — they flatter by default. This lesson is how you get the truth out of it instead: a blunt professional critique, a set of stress tests, and an honest read on how much to trust an AI’s eyes.

The trick: hand it a stranger’s work

The single most reliable way to unlock an honest opinion is to frame the logo as someone else’s. The model has no reason to protect a stranger’s feelings, so the honesty rises sharply. Attach your near-final mark and send this:

A designer sent me this logo for my [type of business] that serves [target customer].
I did NOT make it and I have no ego about it — I need the truth. Score it out of 10 on
each of these, and for every point you take off, say exactly why and what you'd change:
1) Legibility (does it read instantly?)  2) Simplicity (does it work as one small shape?)
3) Memorability/distinctiveness  4) Fit for my business and audience  5) Versatility
(does it survive tiny sizes and one color?). Then give one overall verdict: keep, tweak,
or start over — and defend it.

The score-out-of-10-per-item structure matters as much as the framing: a number demands a reason, so you get “the wordmark is too thin at small sizes” instead of “looks great!”

Those five scores aren’t random — they’re the checklist a pro runs on every mark:

The stress tests a pro runs in their head

Right after the critique, pressure-test the mark the way a professional would:

Now stress-test this same logo the way a professional would. Answer honestly, one line each:
- Does it still work shrunk to a 32px browser favicon?
- Does it still work in pure black-and-white (no color to lean on)?
- Would it look at home next to my competitors, or dated/generic?
- Is it accidentally similar to any well-known logo I should worry about?
- Would you be proud to put this on a business card and a hat? If not, why not?

Turn the critique into your next move

Don’t just read the feedback — make it write your fix. This bridges mentor straight into maker:

Based on the single most important weakness you just identified, write me the exact
image-generation prompt I should paste back into the logo tool to fix it — keeping
everything that already works. Explain in one sentence what your change is meant to
improve so I can judge the result myself.

Paste that prompt back in, generate, and you’ve closed the loop: critique → fix → a better mark, in three messages.

How much to actually trust it

An AI’s eyes are a sharp second opinion, never an oracle. Keep these four rules and you’ll get the value without being led astray:

• Run it twice. Ask for the critique again in a fresh chat and keep only the findings that show up both times — a one-time-only flag is often the model guessing.
• Trust specific, ignore vague. “The wordmark is too thin at small sizes” is worth acting on. “It lacks clarity” is filler — bin it.
• Cap the list. Add “give me only the 3 that matter most” so it doesn’t pad to nineteen problems to look thorough.
• You decide. It produces; you judge. If your gut and the critique disagree and you can say why, your gut wins.

Watch this step

A concrete watch-along of feeding a design to AI and using a structured critique prompt to get honest feedback instead of applause — the exact pattern from this lesson. It’s from an independent creator — credited below, so go give them a follow.