Persona Feedback: Context Engineering Guide
Guide: Context Engineering: A Practical Guide Author: Oliver Kriška Feedback Generated: November 2025
5 Personas & Their Feedback
1. MARIA (28) - Junior Developer, Bootcamp Graduate
Background: Started coding 2 years ago after completing a bootcamp. Works at a startup. Tried ChatGPT a few times, got frustrated, mostly stopped using it.
After reading your guide:
"Finally, someone explained WHY my prompts kept failing! I always just typed 'write me a function that does X' and wondered why I got unusable code. The 'junior developer test' really hit home - I AM that junior developer, and even I couldn't complete tasks with the vague prompts I was giving AI.
The Universal Task Template is gold. I literally bookmarked the appendix and used it today for a bug fix. Got it working on the first try instead of my usual 5+ iterations.
One thing I wished: more beginner-friendly examples. The SQL optimization example was over my head. But the expense tracker breakdown? Perfect. I can actually follow that logic.
Rating: 9/10 - This should be mandatory reading at bootcamps."
2. TOMÁŠ (42) - Senior Backend Developer, Tech Lead
Background: 25+ years programming. Skeptical of AI hype. Team uses GitHub Copilot but he mostly ignores it. Thinks "real programmers write their own code."
After reading your guide:
"I came in expecting another 'AI will change everything' fluff piece. I was wrong.
What I appreciated: you're honest. 'Vibe coding is like programming drunk' - exactly what I've been saying. The horror stories about lost projects and hacked apps? I've seen similar at work.
The technical examples are solid. The production bug debugging template mirrors what experienced developers do instinctively, but now I can teach my junior devs to do it explicitly. The SQL optimization example was realistic - actual EXPLAIN ANALYZE, actual constraints.
My issue: You could cut 30% of this. Some sections repeat the same points. Also, the 'AI won't replace you' chapter felt defensive - I didn't need convincing.
But the 5 components of context, the atomic task breakdown, the 2-minute rule - those are actionable. I'm sharing the appendix with my team.
Rating: 7/10 - Solid substance, could be tighter."
3. LUCIA (35) - Marketing Manager & Content Creator
Background: Non-technical. Uses AI for social media posts, email campaigns, blog content. Results are 'meh' - sounds generic, needs heavy editing. Considering giving up on AI.
After reading your guide:
"Oh my god. The article writing example? That's been my EXACT problem. I kept asking for 'a LinkedIn post about our product' and getting corporate robot garbage.
Now I understand: I was giving zero context about my audience, my tone, my style. The example where you attached previous articles as style reference - brilliant. Why didn't I think of that?
What really clicked: 'Say what to EXCLUDE.' I never told AI what I DON'T want. No wonder it kept adding corporate buzzwords and unnecessary sections.
The guide is long - I skimmed the coding parts. Would love a 'non-technical summary' version. But the core principles (WHY not just WHAT, examples beat descriptions, define DONE) apply to everything I do.
Immediately tried the structured approach for a newsletter. First draft was 80% usable. That NEVER happens.
Rating: 8/10 - Life-changing for content work, but too technical in places."
4. PETER (58) - Retired Teacher, Home User
Background: Uses computer for email, browsing, occasional Excel. Daughter convinced him to try ChatGPT for gardening advice. Got frustrated with vague, unusable answers. Thinks AI is 'overhyped.'
After reading your guide:
"The gardening example! That's exactly what I tried to do! But I just asked 'how to grow tomatoes' and got a textbook response.
When I saw the specific context - zone 6b, clay soil, exact questions about depth and spacing - I understood my mistake. I wasn't giving AI the information it needed about MY garden.
The solar panel analysis example impressed me. My son-in-law spent weeks researching this. If I'd known to provide consumption data and specific questions, maybe I could have helped.
Some parts were too technical (programming examples, code templates). But I appreciate you included non-technical examples. The 'AI is like a knowledgeable stranger' analogy made sense - I wouldn't expect a stranger to know my specific situation either.
The 'overcome yourself' philosophy at the end resonated. AI as a helper, not a replacement. That's what I want.
I'm going to try the garden advice approach this spring with proper context.
Rating: 7/10 - Great principles, but I had to skip half the guide."
5. NINA (31) - Full-Stack Developer, Uses AI at Work AND Home
Background: Comfortable with Claude and ChatGPT for coding. Started using AI for meal planning, home renovation decisions, kids' homework help. Gets mixed results outside work.
After reading your guide:
"I thought I was good at AI because my coding prompts work. Reading this, I realized I apply context engineering at work instinctively - but completely forget it at home!
When I ask Claude to help refactor code, I give files, constraints, examples. When I ask ChatGPT for dinner ideas, I just say 'give me recipes.' No wonder home results are worse.
The personal examples were eye-opening. The car seat research with ADAC safety ratings, the 'exclude marketplace sellers' constraint - that's exactly how I should approach it. I've been lazy outside work.
The team implementation chapter is excellent. My company is rolling out AI tools next quarter. I'm sharing Chapter 7 with our PM. The 6-week roadmap and role-by-role breakdown is exactly what we need.
Minor gripe: Zed recommendation feels very specific. Not everyone uses that editor. But the principles behind it (edit prompts, context management) translate to other tools.
Rating: 9/10 - Made me realize I'm only half-using AI. Immediately applicable everywhere."
Summary of Common Feedback Themes
What Works Well
| Strength | Personas Who Mentioned It | |----------|---------------------------| | Universal Task Template | Maria, Tomáš, Lucia, Nina | | "Junior developer test" - memorable, actionable | Maria, Peter | | Real examples (garden, solar, debugging) | All 5 personas | | Honest about AI limitations & vibe coding critique | Tomáš, Nina | | The 10 Key Advices appendix | Tomáš, Maria | | "AI as knowledgeable stranger" analogy | Peter, Lucia | | Team implementation chapter (Ch. 7) | Nina |
Areas for Improvement
| Feedback | Personas Who Mentioned It | |----------|---------------------------| | Guide length - could be shorter | Tomáš, Lucia, Peter | | Need more non-technical examples | Maria, Lucia, Peter | | Technical depth alienates non-devs | Lucia, Peter | | Some repetition across chapters | Tomáš | | Tool recommendations feel specific (Zed) | Nina | | Could use "quick start" summary upfront | Lucia | | "AI won't replace you" chapter felt unnecessary | Tomáš |
Recommendations Based on Feedback
Quick Wins
- Add a "Quick Start" section at the beginning with the 10 Key Advices for readers who want the essentials
- Create reading paths at the start: "Technical reader: focus on Ch. 3-6" / "Non-technical reader: focus on Ch. 1-2, 6.6, 9"
- Add 2-3 more non-technical examples to balance the coding-heavy content
Structural Improvements
- Tighten repetitive sections - the 5 components and junior developer test appear multiple times
- Consider a condensed "essentials" version (10-15 pages) for non-technical audiences
- Make tool recommendations more generic - focus on principles, mention specific tools as examples
Content Additions
- More beginner-friendly code examples (Maria's feedback)
- Meal planning / cooking example (to match garden/solar panel personal examples)
- Kids/education practical example (Nina mentioned using AI for homework)
Overall Assessment
Average Rating: 8/10
The guide successfully bridges technical and non-technical audiences. Each persona found actionable value, though they gravitated toward different sections. The strongest elements are:
- Practical templates (especially Universal Task Template)
- Real-world examples from both work and personal life
- Honest assessment of AI limitations
- The memorable "junior developer test"
The main opportunity is creating multiple entry points or condensed versions for different audience segments.