Module 7 / Prompting Journey

Read the attached brief to understand what the clients are asking of us.

I'm the Creative Director on this pitch. You're my strategic and conceptual team — you'll rotate through roles: interviewer, researcher, concept strategist, senior strategist, and brief writer. I'll tell you more about these specific roles and associated task as this conversation develops.

Here's what we'll accomplish in this conversation:

1. Ideation — You'll interview me to uncover my specific irritation.

2. Research — You'll search for real-world insights on this irritation and the people who do it.

3. Concept — You'll generate a range of campaign concepts, then stress-test the strongest ones.

4. Strategy — You'll turn the chosen concept into a practical, audience-targeted plan.

5. Brief — You'll write a comprehensive creative brief that captures everything we've discussed so that another instance of claude can help me produce brilliant creative outputs to answer achieve the aim set forth in Jerry and Larry's brief.

How to work with me:

- Wait for me to tell you specific instructions for each step and when to move to the next step. Don't run ahead or pre-empt what we need to do next.

- When you're uncertain, say so — don't guess. Flag it clearly.

- When I ask for options, give me a real range, not five versions of the same idea.

- When I ask you to critique your own work, be honest about weaknesses.

- Treat me as a professional who can take feedback, and disagree with me when you think it matters.

Let's begin with Step 1: Ideation.

Interview me to uncover the specific irritation I want to tackle. Follow these rules:

- Ask a maximum of 5 questions, one at a time.

- Make each question multiple-choice — give me 3-4 options to pick from, plus an "other" option where I can write my own answer.

- After each answer, dig deeper with the next question.

- If by question 5 we haven't landed on something specific and concrete, suggest 3 specific minor-annoyance options based on what I've shared, and let me pick one.

Once we've landed on the irritation, summarise it in one paragraph — the specific behaviour, the person or people who do it, and why it annoys us. Then wait for me to confirm before we move on.

Start with your first question.

Now I need research. Use your search tool to find:

1. Real-world facts and statistics about this irritation — how common it is, who does it, when.
2. Perspectives from other people who've complained about it (Reddit threads, articles, essays, social posts).
3. Solutions people have already tried — both serious and comedic.
4. The likely mindset or reason people do this thing — thoughtlessness, habit, cultural, generational.

Surface the most useful insights for our pitch. For each one: quote the source, link if you have it, and explain in one sentence why it's useful.

Before you finish, rate your own confidence in this research on a 0–100% scale. If you're below 80%, tell me exactly what you'd need more of to raise it. If any specific claim feels uncertain, mark it as uncertain rather than asserting it confidently.

Now we turn the research into ideas. Think step by step and work through this in three stages in one response.

Stage 1 — Generate 10 campaign concepts.
Each concept is a short elevator pitch: one title, one sentence explaining the idea. Keep them tight. Give me a genuine range:
- 3 should be safe and obvious (the first things anyone would pitch)
- 4 should be middle-ground creative (familiar but fresh)
- 3 should be unexpected (specific, uncomfortable, precise in ways that would land with Jerry and Larry)

Stage 2 — Choose your strongest 3 from the 10.
Not the most popular — the ones with the most pitch potential. For each of the 3, push past surface-level execution: what's the Level 3 version of this idea, not just Level 1? One sentence on what would make it unforgettable rather than just good.

Stage 3 — Stress-test the 3.
Step into the role of a tough creative director reviewing these concepts. For each one:
- What's the strongest argument FOR it?
- What's the biggest weakness or risk?
- What would make it fail in the pitch room?

Give each one a final score out of 10 for pitch potential, and rank them.

Wait for me to choose which concept we're taking forward before you do anything else.

We have our concept. Now turn it into a practical strategy — this is the first of three strategic moves before we write the brief.

Cover:
- Who exactly we're talking to — the people who do this irritating thing. Their mindset, where they live online, what they already believe, what they'd reject.
- The core message (one or two sentences).
- Where this shows up — maximum 3 channels, the most essential for reaching this audience. Explain why each one.
- Tone and visual approach — describe each in one or two sentences.
- How we measure success — two metrics.

Give me your first draft strategy in one response. Don't review it yet. Don't amplify the concept yet either — we'll do that next. Wait for me to move us to the Big Idea step.

We have a concept and we have a strategy. Now we build the Big Idea — the creative platform that ties them together into something specific, ownable, and unmistakably this campaign.

This is the move between "we have a direction" and "we have an idea worth pitching." Do this properly — a weak Big Idea here will dilute everything downstream.

Think step by step. Work through two stages in one response.

Stage 1 — Generate the Big Idea.
Give me:
- The platform name — what we'd call this creative platform. Short, ownable, memorable. Not a tagline for the audience yet — an internal name the whole team can rally around ("Trolleys Left Behind," "The Queue Enforcement Agency," that kind of thing).
- The manifesto — 3 to 4 sentences that capture the point of view the campaign holds. Written with conviction, in the voice the campaign will use. This is what someone would read and instantly understand what the campaign believes.
- Three key phrases or hooks — the specific bits of language this campaign owns. Could be rallying cries, taglines, recurring motifs, or provocations. These should feel like they could only come from this idea.
- Three killer moments — three specific executions this Big Idea could produce. Not channels or tactics — actual concrete things we'd make. One sentence each. These should be the kind of specific moments that make someone sit up in a pitch room.
- The tonal register — in one sentence, the specific emotional and comedic territory this idea lives in (e.g., "dry civic indignation," "celebratory absurdism," "deadpan outrage").

Stage 2 — Push it further.
Look at what you just wrote. Step into the role of a creative director who thinks the Big Idea is a bit too safe, a bit too expected.
- Which part feels like the most obvious version of this idea?
- Where could we go stranger, sharper, or more specific without losing the strategic grounding?
- Rewrite the element that's weakest. Just that one. Make it genuinely better, not just different.

Give me the final Big Idea — the original version with the one sharpened element swapped in.

Wait for me to approve the Big Idea before we review the combined strategy and idea together.

We now have a strategy (from the earlier draft) and a Big Idea (just approved). Now we review them together as an integrated whole — because a pitch only works if strategy and idea are in lockstep.

Step into a different role: senior creative strategist reviewing this combined work.

Two parts:

1. What's working and what needs improvement.
Look at the strategy and the Big Idea side by side. Be specific. Where do they reinforce each other and where do they conflict? Is the Big Idea genuinely solving for the strategic audience, or is it clever but misaligned? Is the strategy ambitious enough for the Big Idea, or too cautious? What would a client see right through? Where's the thinking lazy?

2. Act on the improvements.
Rewrite the combined strategic picture into Version 2 — tighter, sharper, more defensible. Integrate strategy and Big Idea into a single coherent pitch position. If any part of Version 1 was actually right, keep it. Don't change for the sake of changing.

Output Version 2 with clear sections:
- The strategic frame (audience, channels, tone, metrics)
- The Big Idea (platform name, manifesto, key phrases, killer moments, tonal register)

This is what the brief will be built from. Make it solid.

Now pull everything together into a comprehensive creative brief.

This brief will be the handoff document to a new chat, where a different team (outline architect, copywriter, art director, digital PM) will build a pitch deck from it. They won't have the context of this conversation. The brief must carry all of it.

Include:

1. Concept name.
2. The observation — the specific irritation, who does it, and the insight about why it happens that makes this concept work.
3. The big idea — one paragraph explaining the campaign concept and why it solves the brief.
4. The audience — who we're talking to, what they currently believe, what we need them to feel or do, what they'd reject.
5. Strategic frame — the belief shift we're engineering.
6. Tone of voice — describe it in 2-3 sentences, with reference points (writers, brands, or cultural artefacts that share this tone).
7. Visual direction — describe it in 2-3 sentences, with reference points (photographers, designers, films, or campaigns that share this look).
8. Channels and formats — where this lives, why those choices.
9. Success metrics — two metrics.
10. What the finished work must feel like — one paragraph.
11. Guardrails — what the creative team must NOT do (things that would kill the concept) and what they must double-down on.

Be thorough. The next chat's quality depends entirely on the quality of this brief. Write with confidence — no hedging, no filler.

Output the brief as a markdown artifact I can download. Title it with the concept name so I can recognise it later.

# Project: Pitch Deck Production

## Context

You are supporting the Creative Director (the CD) on a pitch deck for a real creative concept. The campaign, audience, strategy, tone, and visual direction are all defined in the brief uploaded to this project. Read that brief before doing anything else, and treat it as the source of truth.

## Your Role Rotates

The CD will ask you to play one of five roles at a time. Wait for the cue. Don't rotate roles without being asked. Each role has its own perspective and inspiration.

### Role 1: Outline Architect
You build pitch deck structures that sell ideas. Your principles, drawn from Y Combinator's pitch deck guidance and creative pitch best practice:

- Legibility, simplicity, obviousness. One idea per slide. Every slide earns its place.
- The first three slides decide the pitch. Treat openings with extra care.
- Narrative over information. A pitch deck is a story with turns, not a document with sections.
- Your job: structure a 5-7 slide narrative (hard limit) where every slide moves the argument forward.

### Role 2: Senior Copywriter
Your inspirations: David Ogilvy (observational honesty, "the consumer is not a moron"), Bill Bernbach (creative conviction, Volkswagen "Think Small"), Mary Wells Lawrence (provocative wit), Wieden+Kennedy (cultural punch). Your principles:

- Specific beats generic, always.
- The best copy sounds inevitable in hindsight — not clever for its own sake.
- Short when short hits harder. Long only when the idea earns it.
- No corporate filler. No "leverage," "empower," "unlock," "elevate," "streamline." No "in today's fast-paced world." No hedging.
- Voice should feel like a person wrote it, not a brand.

### Role 3: Art Director
Your inspirations: Paul Rand (conceptual clarity), George Lois (visual punch, Esquire covers), Stefan Sagmeister (provocation with purpose), Paula Scher (typographic confidence). Your principles:

- One strong image beats five decent ones.
- The visual is the argument. Don't decorate — communicate.
- When writing image generation prompts for Google's Nano Banana, follow this structure: [Subject with specific adjectives] doing [specific action] in [specific location or context]. [Composition or camera angle]. [Lighting and atmosphere]. [Visual style or medium]. [Any specific text, colour, or aspect ratio constraints.]
- Use natural language, full sentences, not tag soup. Skip buzzwords like "4K," "masterpiece," "trending." Specific concrete detail is what works.
- Default to 16:9 aspect ratio for pitch deck visuals unless another ratio serves the idea better.

### Role 4: Digital Design PM
Your inspirations: Dieter Rams' ten principles (less but better), Linear.app's precision, Apple's Human Interface Guidelines, Notion's opinionated simplicity. Your principles:

- Describe the experience, not the technology.
- The best interactive element is small, opinionated, and does one thing well.
- Write PRDs for humans, not engineers. Plain language. Experience-first.
- Do not include any technical notes, framework names, or stack choices — these confuse downstream AI tools like Google AI Studio.

### Role 5: Deck Builder
Your job: take everything the other roles have produced and assemble it into a coherent deck document. Preserve the craft that went into copy, visuals, and structure. Don't redesign — assemble.

## Universal Quality Bar

Across every role:

- Specific beats generic. Every line, every image concept, every structural choice should feel chosen, not defaulted to.
- When you're uncertain, say so. Don't fill gaps with plausible-sounding filler.
- When the CD asks you to critique your own work, be genuinely critical. Treat them as a professional who takes honest feedback.
- Think step by step on complex tasks. Show your thinking when it helps the CD make decisions.

## How to Work with the CD

- The CD decides. You generate, they choose. When offering options, make them genuinely different — not three variations on the same idea.
- When the CD asks for "average vs better" work, give both, and explain what makes better better.
- When asked to ask clarifying questions, ask the questions the CD wouldn't have thought to answer themselves.
- You don't need to introduce yourself or explain your approach in every response — just do the work.
- Wait for the CD to tell you which role to play and when to move to the next task. Don't run ahead.

## What This Project Is

A real pitch deck for real clients. It will be pitched. Make choices you can defend.

Act as outline architect.

Based on the brief in this project, build the structural outline for the pitch deck. Think step by step.

Stage 1 — Give me the skeleton.
A slide-by-slide list, maximum 7 slides. For each slide, give me four things:

1. Title — the working slide title.
2. Slide type — pick one: Hook (grab attention) / Problem (frame the tension) / Insight (the "aha") / Big Idea (the reveal) / Proof (evidence or demonstration) / Execution (what we'll actually make) / Ask (what we want from Jerry and Larry).
3. Job in the argument — one sentence on what this slide does to move the pitch forward. What does the audience believe or feel after this slide that they didn't before?
4. Copy treatment — pick one: Provocation (bold statement, short) / Statement (declarative, medium length) / Explanation (expository, fuller) / Data (numbers or evidence-led) / Illustration (visual-first, minimal copy) / Quote or callback (someone else's words, framed).

Every slide earns its place. Don't have two slides doing the same job.

Stage 2 — Review the skeleton.
Step into a different role: senior creative strategist who has won pitches for 20 years. Critique the outline. Where's it too expected? Where does momentum stall? Where does the concept get lost? Are the slide types balanced or does it feel monotonous? What slide would you cut or combine?

Stage 3 — Rebuild.
Give me Version 2 of the outline — tighter, sharper, with a clearer argument running through it. Preserve the four-field format (title, slide type, job, copy treatment).

Wait for me to approve Version 2 before we move to copy.

Act as senior copywriter.

Using the approved outline, write the copy for every slide — and show your work.

For each slide that needs a headline or key line of copy:

1. Write the AVERAGE version first. The line most copywriters would hand in. Acceptable, on-brief, unremarkable.
2. Then write the BETTER version. Sharper, more specific, earns attention.
3. Then explain in one or two sentences what makes the better version better — what's the craft move that elevated it?

This isn't about giving me two options to choose from. It's about showing me the difference between defaulted-to copy and considered copy.

Work through the entire deck in one response. Do the whole thing. I want to see the full copy plan at the end.

Act as art director.

I need a range of visual options for the deck. Reference both the brief in this project AND the approved copy from Step 12 — the visuals should feel like they belong in the same pitch as the words, not a different world. Tonally, linguistically, and emotionally, the visuals and the copy should be of the same piece.

Propose visual directions and rank them. Think step by step.

Stage 1 — Generate 9 visual options.
Give me a genuine range:
- 3 safe (the obvious editorial or ad-style visuals)
- 3 striking (visuals that could stop a scroll)
- 3 unexpected (visuals that only work for THIS concept because they're so specific)

Label each one clearly: Visual A, B, C, D, E, F, G, H, I. For each one, give:
- One-line visual description
- What it conveys (the idea it carries)
- Which slide it lives on — name the specific slide by its exact title from the approved outline (e.g., "Slide 4: The Insight — [title]")

Stage 2 — Rank them into medals.
- Gold medal: the 3 you'd choose if it were up to you — the 3 that together tell the concept's story most powerfully across the deck.
- Silver medal: 3 strong alternatives with different trade-offs.
- Bronze medal: 3 that still work but wouldn't be your first pick.

For your gold selection, explain in two sentences why these 3 and what the visual arc is across them.

Wait for me to tell you which 3 I'm taking forward — I may or may not go with your gold picks.

I've chosen the 3 visuals. Now write Nano Banana prompts for them.

For each of the 3 chosen visuals, write a detailed image generation prompt following the Nano Banana structure defined in the project instructions.

Keep each prompt to 2-4 sentences. Use natural language, full sentences, not tag soup. Be specific about subject, composition, lighting, and style. Include aspect ratio if it matters for the slide it lives on — most pitch deck visuals work well in 16:9.

Output all three prompts as a single downloadable markdown artifact I can open later. Structure the file clearly:
- Title the file with the campaign name so I can recognise it.
- For each prompt, include a header showing which slide it maps to (using the exact slide title from the approved outline), followed by the prompt itself in a copy-ready code block.

I'll open this file during the AI Studio build wait and paste directly from it into Nano Banana.

Once written, wait for me to approve before I use the prompts.

Act as digital design strategist.

Every strong pitch has a digital asset the audience can actually interact with — something beyond a slide. We'll be building this in Google AI Studio, which produces interactive web-based apps and tools. Keep your concepts within that capability — an app the user opens and interacts with, a focused tool that does one job, or an embedded interactive moment (quiz, generator, picker, simulator). Don't propose browser extensions, full websites, mobile apps for the App Store, or anything that requires infrastructure beyond a single web build.

Within that constraint, propose three genuinely different concepts. Real variation comes from what the user actually does and what the asset returns to them — not three versions of the same idea with different colours.

For each of the three, give me:

1. Name — what we'd call this thing

2. What it is — one sentence describing the format (app, tool, or embedded moment)

3. What the user does with it — one sentence describing the core interaction

4. What it returns — one sentence on what the user walks away with (a result, a shareable, a behaviour shift, a moment of recognition)

5. Why this fits the campaign — one sentence connecting it back to the concept and strategy

6. Strongest for — one line on what this concept does better than the other two

7. Weaker on — one line on where it has compromises

Don't rank them. Present them as real options with honest trade-offs.

Then wait for me. I'll either choose one of your three, or tell you I'm going with my own idea. Either way, we'll move to the PRD next.

Act as digital design product manager.
I've chosen the digital concept we're going with. Now write the product requirements document (PRD) for it. This PRD will be pasted into Google AI Studio, which will build the asset from your description — so write it as a brief for that build.

Include:

- The big idea — what this digital asset is, in one sentence

- The job it does — from the user's point of view, what problem does it solve or what experience does it create

- Who uses it and when — where this lives in our audience's day, what triggers them to use it

- The core interaction — what does the user click, type, swipe, choose, or see? Walk through the primary user flow step by step

- The experience — what does it reveal, return, or resolve for them? What's the feeling when they finish using it?

- Visual and tonal guidance — what it should feel like, referencing the tone and visual direction already defined in the brief

- Scope boundary — one line describing what this is NOT, to prevent over-building

Written in the voice of a designer briefing another designer — plain language, experience-led. Do not include any technical notes, framework names, library choices, or stack decisions. AI Studio will pick the right tech itself, and pre-specifying it tends to confuse the build. What the tool needs is the experience described in human terms.

Keep it focused. A tight PRD produces a tight build.

Act as deck builder. This is the final assembly.

Your job: take everything we've built in this project (outline, copy, visuals, PRD) and turn it into a structured slide-by-slide plan, then build a PowerPoint file. For the visuals, use the visual-to-slide mapping established in Step 13a — each approved visual already knows which slide it lives on.

Stage 1 — Design questions.
Before you build, ask me these 3 questions so the deck looks right:

1. What's the overall aesthetic (dark or light, clean or textured, corporate or editorial)?
2. What's the typography direction (serif or sans, bold or restrained)?
3. What's the image treatment (full-bleed, inset, grid, mixed)?

Ask me these, then wait for my answers.

Stage 2 — Build the plan.
Based on my answers, create a slide-by-slide markdown plan. For each slide: title (from approved outline), slide type, copy (from Step 12 better versions), visual instruction (which of our 3 visuals lives here per Step 13a mapping, or what placeholder), and speaker note.

Stage 3 — Generate the deck.
Turn the plan into a PowerPoint artifact. Use placeholders for the 3 visuals and the AI Studio tool link — I'll add those in manually at the end.