Purpose
Turn a set of design decisions and screens into a tight, persuasive spoken narrative that gets alignment instead of endless critique — sequenced so the audience understands why each decision was made before they see the pixels, and calibrated to what an executive, product, or engineering audience actually cares about.
When to Use
Invoke this skill when the user asks to:
- Prepare a design review or design critique presentation
- Present design work to executives or leadership for a decision or sign-off
- Build a narrative or talk track for a stakeholder alignment meeting
- Explain the reasoning behind design decisions to cross-functional partners
- Turn a set of screens or Figma frames into a story instead of a slide dump
- Handle expected objections or pushback on a design direction before the meeting
Inputs Required
Before starting, collect:
- Design work: What was designed — the feature, flow, or product area, plus the number and rough content of screens/artifacts to walk through
- Audience type: Executive (leadership/VPs) / Product (PMs, product leads) / Engineering (eng managers, tech leads) — or a mix, with the primary decision-maker named
- The problem: The user or business problem this design solves
- Key decisions: 2–5 specific design decisions the user made and wants to defend (e.g., "chose a single-column flow over a wizard")
- The ask: What the user needs from the room — approval, resourcing, feedback on a specific fork, or a go/no-go decision
- Constraints and tradeoffs: What was ruled out, deprioritized, or deferred, and why
- Known tension: Any anticipated objection, skeptic in the room, or prior disagreement
If audience type is missing, default to a mixed product-and-engineering audience and note the assumption. If the ask is missing, ask one clarifying question — the ask determines the entire narrative shape.
Instructions
Step 1: Calibrate to the Audience
Set the narrative lens before writing a word of script. Each audience weighs decisions differently:
- Executive: Leads with business impact and risk. Wants the problem sized, the bet named, and the decision framed as a tradeoff between outcomes. Minimize craft detail; maximize "why now" and "what it costs us to not do this." Talk in metrics, timelines, and strategic fit.
- Product: Leads with the user problem and the solution logic. Wants to see the alternatives considered and why this path serves the roadmap. Show the exploration; connect each decision to a user need or PRD requirement.
- Engineering: Leads with feasibility, scope, and edge cases. Wants to know what's fixed vs. flexible, where complexity lives, and what the design assumes about the system. Flag states, error paths, and effort-vs-value tradeoffs explicitly.
Pick the primary lens; note where a mixed audience needs a bridging sentence for a secondary group.
Step 2: Build the Narrative Spine
Structure the story on a five-beat arc. This is the backbone every subsequent step attaches to:
- Problem — the tension the audience already feels or should feel
- Insight — the non-obvious thing learned that reframes the problem
- Exploration — the space considered and the paths not taken
- Decision — the chosen direction and the reasoning that makes it defensible
- Impact — what changes for the user and the business if this ships
Write a one-sentence version of each beat first. If the five sentences don't hold together as a story on their own, the script won't either — revise the spine before expanding.
Step 3: Write the Opening Hook
Draft 2 opening options (each 3–4 sentences, ~30 seconds spoken):
- Option A — Stakes-first: Open on the cost of the current state or the size of the opportunity. Best for executive rooms.
- Option B — Moment-first: Open on a specific user moment or a sharp observation from research. Best for product rooms. Each hook must land the problem in the audience's terms and end by signaling where the talk is going — no "today I'm going to walk you through my designs."
Step 4: Write the Per-Artifact Talk Track
For each screen or artifact the user is presenting, write a talk-track block containing:
- Setup line: One sentence stating what decision or moment this screen represents (the "why you're seeing this") — spoken before the audience studies the pixels
- Walk line(s): 1–2 sentences guiding attention to the specific element that matters, not a description of the whole screen
- Decision tie-in: One sentence connecting this artifact back to a key decision or the audience's priority from Step 1
- Transition: A half-sentence that carries momentum into the next artifact
Keep each block to spoken length (~20–40 seconds). Sequence artifacts to follow the narrative spine, not the Figma file order.
Step 5: Pre-empt Objections
For each anticipated objection (minimum 3, including the known tension from inputs), write:
- The likely objection in the audience's own words
- The pre-empt or response — a 1–2 sentence answer that acknowledges the concern and reframes it around the decision's reasoning or a tradeoff already made
- Placement: Whether to address it proactively in the script or hold it in reserve as a Q&A answer
Step 6: Write the Closing and the Ask
Draft the closing (~30–45 seconds):
- Restate the impact in one line
- State the explicit ask from inputs — make it a single, specific, answerable request
- If approval is the ask, name what a "yes" unblocks and the timeline End with a direct handoff to the room, not a trailing "so… any questions?"
Step 7: Assemble the Full Script Document
Compile everything in presentation order with speaker-note formatting, plus a timing estimate for the full talk.
Output Format
- Structure: A speaker-note script in presentation order — Audience lens summary → Narrative spine (5 beats) → Opening hook (2 options) → Per-artifact talk track (one block per screen) → Objection pre-empts table → Closing and ask → Total timing estimate
- Length: 900–1,400 words depending on artifact count
- Format: Structured document; talk-track blocks as labeled sub-sections; objection pre-empts as a 3-column table (Objection / Response / Placement)
- Sections: Audience Lens → Narrative Spine → Opening Hook → Talk Track → Objection Pre-empts → Closing & Ask → Timing
- Tone: Spoken, confident, first-person — script the words the presenter actually says out loud, not slide copy or a written report. Concise; no filler transitions.
Quality Checklist
Before delivering output, verify:
- [ ] Every key decision from the inputs is defended somewhere in the talk track (count them against the input list)
- [ ] The narrative follows the five-beat spine — problem and insight land before any screen is described
- [ ] Each per-artifact block includes a setup line stated before the walk line, not after
- [ ] At least 3 objections are pre-empted, and the known tension from inputs is one of them
- [ ] The closing states one specific, answerable ask — not "let me know your thoughts"
- [ ] The audience lens is applied consistently — executive scripts don't drown in craft detail; engineering scripts flag states and feasibility
Notes
- The most common failure in design presentations is describing screens before establishing why they exist — the setup-line-before-walk-line rule fixes this and is the single highest-leverage habit here.
- For a mixed room, write to the primary decision-maker and add one bridging sentence per secondary group rather than trying to satisfy everyone in every beat.
- If the user has more than ~8 artifacts, recommend grouping them into 3–4 narrative moments rather than walking every screen — a story has scenes, not a slide-per-tap tour.
- Encourage the user to read the script aloud once against the clock; the timing estimate assumes ~130 words per minute of speaking.
How to use this skill
This file is a Claude Code skill in Anthropic's standard format — markdown with YAML frontmatter at the top.
Installing in Claude Code:
~/.claude/skills/<folder-name>/SKILL.md— auto-loads in every session anywhere on your machine (recommended for skills you'll use across multiple projects)<your-project>/.claude/skills/<folder-name>/SKILL.md— auto-loads only when working in that project (recommended for project-specific skills)
Restart Claude Code or open a new session to register.
Using in Claude.ai, ChatGPT, Gemini, or other AI tools:
- Drag this file into a new chat, or copy and paste its contents
- Send your task or request as your next message — the AI will apply the skill where it fits
For Claude Code skill details: https://docs.anthropic.com/en/docs/claude-code/skills