3 AI Design Tips (Beginner-Friendly) You Can Use Today

Stop guessing. Start creating. Here are 3 simple moves that instantly level up your AI design output.

Author

Andrew Hazen
February 02, 2026

AI Design University — Weekly Drop 🚀

Top 3 AI Design Tips for Beginners (That Actually Work)

If you’re new to AI design, you don’t need more tools.
You need better inputs, tighter constraints, and a repeatable loop.

Here are the 3 tips I’d tattoo on every beginner’s workflow.

Tip #1 — Use a “Prompt Recipe,” Not Random Prompts

Most beginners type vibes. Pros use a structure.

Steal this prompt recipe (copy/paste):
Goal: what you want
Audience: who it’s for
Style references: 2–3 specific influences
Constraints: format, size, colors, do/don’t
Output: what you want returned (variants, JSON, steps, etc.)

Example (beginner-safe):

Goal: Create a clean hero banner for AI Design University
Audience: beginners learning AI design
Style: modern, minimal, high-contrast, “tech university” vibe
Constraints: 16:9, large headline space, 2 accent colors max, no clutter
Output: 3 variations + short headline options

Why it works: your AI can’t read your mind—so give it a box to win inside.

Tip #2 — Build a “Constraint Stack” Before You Generate Anything

Constraints aren’t limitations. They’re quality control.

Use this fast constraint stack:

  • Format: (IG post, 16:9 hero, logo, flyer, etc.)

  • Hierarchy: what must be seen first, second, third

  • Palette: 2–3 colors max

  • Typography: 1–2 font styles max

  • Do/Don’t List: (do: clean, bold; don’t: busy, tiny text, 12 icons)

Beginner move that makes you look advanced:
Write a Do/Don’t list and include it in every prompt.

Example:

  • Do: bold headline, clear spacing, minimal elements

  • Don’t: gradients everywhere, tiny body text, overly complex backgrounds

Tip #3 — Run the “3-Pass Loop” (Generate → Critique → Regenerate)

Beginners generate once and hope.
Pros iterate like machines.

Pass 1: Generate
Ask for 3–6 variations, not one “perfect” output.

Pass 2: Critique (this is the cheat code)
Paste this exact critique prompt:

“Critique this design like a creative director. List the top 5 improvements based on: clarity, hierarchy, contrast, alignment, and simplicity. Then suggest 3 alternative directions.”

Pass 3: Regenerate with direction
Pick one direction, then tighten constraints.
Rinse repeat.

Rule: Don’t do 20 random iterations. Do 3 disciplined iterations.

Your 5-Minute Challenge (Do This Today) âś…

Pick one: a logo concept, a social post, a landing hero, a menu header—anything.

  1. Write a Prompt Recipe

  2. Add a Constraint Stack

  3. Run the 3-Pass Loop

Reply to this email with what you made and the prompt you used—
and we’ll feature a few in the next issue.

Did You Know?

AI design facts/stats you most likely did not know:

  1. Designers feel the efficiency boost… but don’t trust the output.
    In a survey of 2,500 Figma users, 78% agreed AI “significantly enhances” their efficiency — but only 32% said they can rely on AI output in their work. That gap is the whole game in AI design right now: speed is here, trust is the bottleneck.

  2. Most designers still aren’t using AI for “core design work.”
    In that same report, 59% of developers use AI for core responsibilities like code generation, but only 31% of designers use AI in core design work like asset generation. And designers report lower quality gains (54%) vs developers (67%).

  3. One creative AI system alone has generated “internet-scale” output.
    Adobe reports its Firefly generative AI has produced 24+ billion assets since launch (March 2023) — a mind-bending indicator of how fast AI-assisted design content is scaling.

—
AI Design University
Learn it. Build it. Design it.