Week 1 · Internal Recap

Week 1 TL;DR

Week 1 was more about setup and workflow than fully testing prototype-first. Here's what we learned.

By the Numbers

📊 Quant Snapshot

6 weekly reflections submitted.

Prototype-first effectiveness
2.5 / 5
Clustered around 2–3
AI comfort
3.17 / 5
Mostly 3s
Effectiveness distribution (1–5 scale)
AI comfort distribution (1–5 scale)
Interpretation: Teams appear more comfortable with the AI tools than convinced by the workflow, suggesting environment and process issues are the bigger blocker — not mindset or AI familiarity.
Qualitative Analysis

🧠 Key Themes from Week 1

Seven consistent themes emerged. Themes 4 and 5 are especially important.

⚙️

Setup & Environment Friction Dominated

  • Debugging, access (GitHub/Vercel), and unclear setup paths slowed progress
  • Time spent "getting things working" exceeded time actually building
Week 1 largely reflects onboarding friction, not the true value of the method.
⚡

Iteration Gets Faster Once Unblocked

  • Strong signal: faster idea → feedback → iteration loops
  • Easier to collaborate with design and evolve concepts quickly
Prototype-first seems strongest for mid-process iteration, not initial setup.
🔁

Net-New Prototyping > Recreating Existing Work

  • Rebuilding flows from Figma/product was slow and frustrating
  • Screen-by-screen or net-new ideas worked much better
This approach currently favors exploration over translation.
🧾

Requirements Still Matter (More Than Expected)

  • Lack of clear written context caused confusion and back-and-forth
  • Prototypes exposed gaps normally filled in by human context
Prototype-first doesn't remove docs — it requires clearer, more explicit inputs.
💬

Feedback Quality Changed with Prototypes

  • Live prototypes → better feedback on interaction and "feel"
  • But harder for self-paced exploration, async comments, and system critique
Prototype ≠ Figma replacement. They surface different kinds of feedback.
🧩

Role Boundaries Blurred (Positive Signal)

  • People working outside typical roles: terminal, PRDs, debugging, etc.
  • More integrated thinking across design / research / product
Early signal of deeper collaboration, not just new tools.
❓

Workflow Still Unclear

  • Open questions: one shared prototype vs. multiple? How to share? How to give feedback?
  • Teams unclear on the operating model and expectations
Need a clearer operating model and workflow scaffolding.
Trade-offs

⚖️ Key Tensions

Four tensions surfaced consistently across teams in Week 1.

Speed
⚡
Setup
🧱
Slow upfront investment, but faster iteration once things are working.
Learning
🧠
Shipping
🚢
Significant time spent learning tools vs. actually building product.
Flexibility
🎨
Clarity
🧭
Powerful and flexible tools, but unclear workflows and expectations.
Prototype
📱
Figma
📐
Interaction feel and realism vs. structural overview and async feedback.
Feedback Loop

📝 In-the-Moment Feedback

Collected via structured forms, Slack links, and emoji tracking.

✅

What's working

  • Logging process is being used (Slack + form + emoji tracking)
  • Good early signal density across the team
⚠️

What we're seeing

  • Feedback is mostly operational friction — expected at this stage
  • Limited product-level insights so far; consistent with where the pilot is
💡

High-value signals

  • Setup + access (Vercel, GitHub) need more support
  • Folder structure and environment clarity matter a lot
  • Large prompts fail; smaller steps work better
  • Need guardrails to avoid modifying shared libraries

Playbook Candidates

These high-value signals are strong candidates for the pilot playbook.

Artifact Comparison

🧪 Prototype vs Figma — Feedback Differences

What changes when the artifact you're reviewing is a live prototype?

📱 Live Prototype

  • ✅ Strong for interaction feel and realism
  • ✅ Easier to react to motion and flow
  • ✅ Research insights applied directly into the experience
  • ❌ Harder to orient, pause, and self-inspect
  • ❌ Harder for async comments and annotations
  • ❌ Difficult to compare branches or system patterns

📐 Figma

  • ✅ Strong for structure, comparison, system thinking
  • ✅ Easier for async comments and self-pacing
  • ✅ Clear orientation across screens and flows
  • ❌ Less realistic interaction feel
  • ❌ Harder to convey motion and live behavior
We may need a hybrid feedback model — not a full replacement of one tool with the other.
Action Plan

🚀 Week 2 Focus

Five priorities heading into Week 2.

1
🧰

Reduce Setup Friction Top Priority

  • Create clear onboarding steps and a known-issues / common-fixes doc
  • Ensure access to Vercel, repos, and environments before Week 2 starts
2
✂️

Encourage Smaller, Iterative Prompts

  • Screen-by-screen or tightly scoped asks significantly outperform big prompts
  • Lean into iteration, not one-shot generation
3
🧭

Clarify the Workflow

  • Define what prototype-first replaces and what it doesn't
  • Clarify when to prototype vs. write vs. align
4
📦

Improve Feedback Setup

  • Share prototype + light written context + optional overview artifact
  • Not a prototype alone — give reviewers enough to orient themselves
5
📊

Continue Capturing Individual Signals

  • Reflections are surfacing friction, role shifts, and comfort levels
  • Keep this feedback loop tight — it's providing useful signal