AI Tips Thursday: 5 Ways to Engineer Smarter Prompts
- Noemi Kaminski
- Nov 1
- 2 min read
Most people write prompts like orders.
Professionals design contexts — because AI doesn’t just respond to words, it responds to frames of thought. Here’s how to shape that thought process 👇
1️⃣ Frame the AI’s mental model
Before asking for output, define how you want the model to think.
You’re not giving it a task — you’re assigning it a mindset.
“You’re a senior UX designer mentoring a junior colleague.”
Now it reasons like a teacher, not a typist.
Framing changes how it processes information, not just what it produces.
2️⃣ Anchor with reference language
Adjectives like “friendly” or “professional” are vague.
Instead, give it a linguistic anchor — a tone sample, sentence, or brand paragraph.
“Match the tone of this excerpt: [paste text].”
Anchors teach through example, not instruction.
They make tone, rhythm, and pacing instantly consistent across long projects.
3️⃣ Chain your constraints, don’t stack them
Most people cram all rules into one line:
“Write 300 words, no jargon, 3 bullets.”
AI handles processes better than laundry lists.
“First, outline 3 bullet ideas. Then expand each in 100 words using plain English.”
You’re giving it sequence logic — like how humans follow steps — which boosts clarity and structure.
4️⃣ Use negative context deliberately
Creativity isn’t just about what to include — it’s about what to exclude.
“Avoid emotional appeals or exclamation marks.”
Negative context trims noise and over-enthusiasm, especially in UX writing or product copy.
Think of it as pruning a tree: fewer branches, stronger shape.
5️⃣ Debug with lens switching
When a response feels off, don’t rewrite the prompt — change the lens.
Ask the AI to view it differently:
“Rewrite as if speaking to a skeptical investor.” (Tone lens)
“Simplify for a 10-year-old.” (Audience lens)
“Explain your reasoning step by step.” (Logic lens)
Switching lenses turns trial-and-error into controlled refinement.
These five steps come straight from my course The Context Advantage: Engineering Better AI Outputs
— where I teach you to design systems for consistent, cross-modal AI performance across writing, marketing, design, and strategy.
🎓 Learn the full 5-element framework: Role, Task, Constraints, Anchors, Exclusions, and more!
🧠 Want weekly deep-dives and frameworks like this?
Join my Newsletter for free resources & workflow breakdowns:
Comments