Design — without being a designer.
Claude can design. Brand sheets, slide decks, landing-page mockups, social ad sets, ad creative — work that would have been a six-month project for a junior designer a decade ago. This module teaches working professionals to use those capabilities without becoming designers themselves.
The trick isn't learning design theory. It's learning to brief design work the way a sharp creative director would. You don't need to know Figma. You don't need to know color theory. You need to know what good looks like for your specific situation and how to describe it precisely enough that Claude can produce it.
What you'll walk out with
A clear mental model of what Claude can and can't design today. The "design RCTC" — a briefing template for visual work. Three end-to-end walkthroughs (brand sheet, board slide deck, landing-page mockup). The iteration discipline that gets you from "draft" to "ready" in three rounds instead of fifteen. And a framework for working with real designers when you need to scale beyond what Claude alone can do.
What this module covers
- What Claude can design today — and what's still better with a human
- The design RCTC — the brief template for visual work
- Walkthrough — Brand sheet from prose (logo direction, palette, type)
- Walkthrough — Board-ready slide deck from a prose outline
- Walkthrough — Landing-page mockup that designers can actually build from
- Iteration discipline — how to redirect without rewriting from scratch
- Working with real designers — when to bring in a human
What Claude can design today.
As of 2026, Claude's design capabilities cover most of what working professionals need for non-publishing work. The honest map of capabilities — and the honest map of where humans still win.
What Claude does well
Layout and information design
Structuring content visually. Where the headline goes, where the supporting points go, how to use white space, when to use a grid. Claude produces clean, on-brand layouts reliably — the kind of work that consumes hours of a designer's time but isn't where the creative thinking lives.
Brand systems
Color palettes, typography pairings, basic logo direction, brand voice extension. Claude can produce a workable brand system from a prose description of your company. The result isn't going to win awards — but it's better than what most small companies have, and it's done in 20 minutes.
Document and slide design
Board memos, investor decks, sales decks, internal presentations. Claude can take a prose outline and produce structured, on-brand slides — text, layout, color, hierarchy. Not "designer-level polish" but "presentable to executives" is well within reach.
Static web pages and mockups
Landing pages, one-pagers, sign-up forms, simple marketing sites. Claude can produce working HTML/CSS that runs in a browser. Combined with the Code module, you can ship actual pages, not just mockups.
Social and ad creative (structural)
Instagram Reel storyboards, ad copy variants, A/B test direction, social post templates. Claude produces the structural work; you (or a tool like Canva) handle the final assembly. The Instagram ad campaign you walked through earlier was 80% Claude design work.
What's still better with a human
- Original logo design — Claude produces workable directions; an actual designer produces the version that becomes your identity
- High-end brand identity — if your brand is a competitive moat, hire a real brand studio
- Custom illustration and original imagery — Claude generates competent placeholders; a real illustrator creates something distinctive
- Print production at scale — Claude doesn't handle bleed, color profiles, and printer specs the way a production designer does
- Anything where the design is the product — if you're a design-led company, design with designers
The honest reframe
Claude is a senior production designer with 10 years of experience. Excellent at making things look polished and on-brand. Not where you go for original creative direction. For 80% of working-professional design work, that's exactly what you need.
The design RCTC.
RCTC works for prompting. It also works for design briefs — Role, Context, Task, Constraints. The framework is identical; the specifics shift toward visual outcomes.
Role
"You are a senior brand designer for a fintech startup." "You are a presentation designer who works with board-level executives." "You are an Instagram-native creative working for small consumer brands." The role steers tone, polish level, and aesthetic.
Context
The brand: company name, what you do, who your customers are, what makes you different. The audience: who'll see this design, what they care about. The use case: where this lives — board room, Instagram, landing page, sales conversation.
Task
The specific deliverable, with format. "A 5-slide board deck." "A 1080x1920 vertical Instagram ad mockup." "A one-page brand sheet with logo direction, color palette, and typography pairing." Be precise about format — the right format determines half the design.
Constraints
The brand colors. The typography preferences. References — "in the style of Stripe's brand language" or "minimal like Linear, not maximalist like Slack." Banned aesthetic moves — "no stock photos of diverse people in business attire." Word counts, slide counts, image counts.
The reference move
The single highest-leverage move in design briefs is naming reference brands or sites that capture the aesthetic you want. Claude has seen most well-known brand systems and can pattern-match.
Examples that work:
- "Color palette in the spirit of Stripe — confident, clean, slightly cool" (vs. "modern colors")
- "Typography pairing like Linear's — minimal sans-serif throughout, no decorative elements" (vs. "good typography")
- "Landing page layout inspired by Notion's — large hero, three feature blocks, social proof, single CTA" (vs. "well-designed landing page")
The reference rule
If you can't name a brand or product whose aesthetic you'd want to borrow, your design brief is incomplete. Spend five minutes finding one — pinterest, mobbin.com, brandnew.underconsideration.com. Then name it in your brief. One good reference is worth ten paragraphs of description.
Brand sheet from prose.
A one-page brand direction — logo concept, color palette, typography pairing — built from a paragraph of company description. The kind of thing you'd otherwise pay $1,500–$5,000 for from a freelance brand designer.
The setup
Pretend you're launching a small business or side project. You need basic brand direction to start designing your website, your business card, your Instagram presence. Not a full brand system — just enough to look intentional rather than improvised.
The brief
ROLE: You are a senior brand designer for early-stage companies. You produce confident, modern brand direction that small companies can execute on without a design team. CONTEXT: - Company name: Riverbend Coffee - What we do: Small-batch specialty roastery and three café locations in a Pacific Northwest mid-sized city - Customers: People who care about coffee — knowledgeable, particular, but not snobby. Mostly 28–55, mix of remote workers and professionals - What makes us different: We source directly from a small set of farms we've visited personally. The story matters as much as the cup. - Brand feel: Warm but professional. Confident without being precious. We want to feel like a real business, not a hobby project. TASK: Produce a one-page brand sheet covering: 1. Three logo direction options — a one-paragraph description of each, no actual logo image needed 2. A primary color palette (3 colors) with hex codes and a one-sentence rationale for each 3. A typography pairing recommendation — heading font + body font, with rationale 4. Three "voice principles" — one-sentence rules about how we sound CONSTRAINTS: - Reference brands I want to feel adjacent to: Stumptown (heritage but modern), Sightglass (clean, deliberate), Counter Culture (knowledgeable, accessible) - AVOID: rustic-mason-jar Americana aesthetic, anything that looks like a 2015 startup, anything overly trendy - Output as clean Markdown with headers — I'll format the final version myself in Canva
What Claude produces
Within seconds, you get back a structured one-page brand sheet. Three distinct logo directions (a wordmark, a roasting-mark icon, a more abstract option). Three colors with proper hex codes and reasoning ("deep brown #2A1810 — anchors the brand in coffee without being literal; warm cream #F5EDE0 — backdrop that feels like the inside of a well-lit café; a single brand accent #C1561F — orange-red, signals the moment of roasting"). Type pairing (a serif for headlines, sans-serif for body — both with specific font names). Three voice principles in plain English.
The iteration
Read it critically. Some things will be right; some won't. If the logo direction "wordmark with a small flame icon" doesn't excite you, say so — Claude regenerates that piece without rewriting the whole document. If the brown is too dark for digital screens, ask for an alternative. Each round is 30 seconds. Four rounds and you have something you'd actually use.
What this gives you
A coherent brand direction that's specific enough to act on. You take this into Canva, Figma, or a real designer's hands and you've already done the strategic work. Twenty minutes of Claude time replaces $2,000 of discovery work.
Board-ready slide deck from a prose outline.
A 10-slide board deck — structure, content, design direction — from a one-page memo. The version that used to require pulling a designer onto your team for a week.
The setup
You're presenting Q3 results to your board next week. You have notes, you have the financial numbers, you have a few key wins and a few honest challenges to address. What you don't have is time to make a polished deck.
The brief
ROLE: You are a presentation designer who builds board-level decks for B2B SaaS executives. You write the deck content AND specify how each slide should look. CONTEXT: - Company: B2B SaaS, 90 employees, $14M ARR - Audience: 5-person board, mix of investors and operators, sharp on numbers - Tone: Direct, honest, no spin. Bad news named in the same paragraph as the recommendation. PROSE INPUT (what I'd say if I were just talking): "Q3 was decent but not what we projected. Revenue $3.6M vs $3.9M plan — miss driven by two enterprise deals slipping to Q4. Gross margin held at 76%. New logos: 14 vs target 18. NRR remained at 118% which I'm proud of. Pipeline coverage going into Q4 is 3.1x, healthy. The real story: we underestimated the procurement cycle on enterprise deals, our average days-to-close grew from 47 to 68. We're adjusting comp plans and the sales process. Q4 needs to land $4.2M to hit annual plan." TASK: Produce a 10-slide board deck. For each slide, give me: - Slide number and title - The key content (text, numbers, sub-points) - A description of the visual layout (where the headline is, where charts go, what color emphasis) CONSTRAINTS: - Brand colors: deep navy #0F2840, warm gold #F0AB00, ink #1a1a1a, paper #fafaf9 - Open Sans throughout - No stock photos. Use simple icons or abstract data visualizations. - Lead each slide with the takeaway — board members read tops of slides first - Maximum 7 elements per slide. No walls of bullets. - Output as Markdown with each slide as a heading block
What Claude produces
A complete 10-slide outline. Slide 1: Q3 Headline with single big number and one-sentence summary. Slide 2: Performance vs Plan — color-coded by metric. Slides 3–4: The wins (specifics, not abstractions). Slides 5–6: The honest challenges with named root causes. Slide 7: What we're doing about it. Slides 8–9: Q4 outlook and pipeline. Slide 10: Specific asks of the board.
Each slide includes the actual text, the visual layout description, and any charts/diagrams it suggests. You can take this directly into Google Slides, Keynote, or PowerPoint and build the actual deck in 30 minutes — most of the thinking is done.
The deeper move
Once you have the outline, ask Claude to produce speaker notes for each slide — what to say while presenting, what to anticipate from the board, how to handle the hardest questions. You now have a deck AND a prep doc. The full board prep that used to take a full day takes two hours.
Landing-page mockup designers can build from.
A complete landing-page mockup — section by section, with copy and layout — that a real designer can take into Figma and produce a finished page from in a day. The pre-work that usually requires a brand strategist and a copywriter.
The setup
You're launching a new product or service. You need a landing page. You can either (a) hire an agency for $5,000–$15,000 and 4–6 weeks, or (b) brief a freelance designer for $2,000–$5,000 and 1–2 weeks, or (c) build it yourself with Claude Code in an afternoon.
Option (c) requires good upstream work — knowing what every section says, what the hierarchy is, what each visual element does. That upstream work is exactly what this walkthrough produces.
The brief
ROLE: You are a landing-page strategist who briefs designers. You produce section-by-section layouts with finalized copy and clear visual direction. CONTEXT: - Product: a 4-week online cohort program teaching small business owners how to use AI for their operations - Audience: 100% owner-operators, 35–60 years old, run businesses with $500K–$5M annual revenue, technical-comfort medium-low - Goal of the page: convince the visitor to sign up for the free preview week - Brand: warm and confident — feels like a knowledgeable friend, not a tech company - Reference brands I'd want to feel adjacent to: ConvertKit's site (warmth, practical), Lenny's Newsletter (smart, no fluff) TASK: Produce a complete landing-page brief, section by section. For each section, give me: - The section name and purpose - The headline (specific copy, not placeholder) - The supporting copy (specific, not placeholder) - The visual layout (where everything goes, what images/icons appear, color emphasis) - The CTA (if present) INCLUDE THESE SECTIONS: 1. Hero 2. Problem the visitor recognizes 3. The solution / what the cohort actually delivers 4. Social proof (use placeholders for testimonials — note where they go) 5. What's included (week-by-week) 6. About the instructor (placeholder) 7. Pricing / Free preview offer 8. FAQ (5 questions) 9. Final CTA CONSTRAINTS: - Total page should feel ~6 scrolls on desktop - Mobile-first thinking (most signups will come from phones) - Brand colors: forest green #2B5D3C, warm white #FBF8F2, charcoal #2A2A2A, accent terracotta #C56142 - Open Sans for body, a serif (Tiempos or PT Serif) for headlines - Output as a structured Markdown document — section headers + clearly labeled subsections
What you get back
A 1,500-word document that fully specifies the page. Every headline is final copy. Every section knows where it sits on the page. Every visual element has a purpose. The designer who builds this can produce a working page in a day without making strategic decisions on your behalf — which is the expensive part.
The handoff
You can do one of three things with this output:
- Hand it to a freelance designer — they build it in Figma + a web framework. ~1 week, ~$2,000.
- Hand it to Claude Code (from the previous module) — Claude writes the actual HTML/CSS. ~30 minutes, free.
- Use it as a Canva spec — build a static visual mockup yourself based on the layouts described. ~2 hours, free.
The leverage
The strategic work — knowing what the headline says, what comes after, what to feature — used to be the expensive part of landing-page design. Claude collapses it from a week of discovery calls and copy revisions to twenty minutes of briefing. The work itself shifts from "make decisions about the page" to "execute decisions already made."
How to redirect without rewriting.
The biggest mistake in design iteration: when something's wrong, people start over. They rewrite the brief from scratch, hoping the next pass nails it. Claude isn't a one-shot tool. Treating it like one wastes the workflow's biggest advantage.
The three-round rule
Most design work should reach "ready" in 3 rounds with Claude. Not 1, not 10. Three.
- Round 1: First draft based on your brief. 60–70% of the way there. Notice what works and what doesn't.
- Round 2: Targeted fixes on the wrong parts. The right parts stay. You're surgical, not destructive.
- Round 3: Polish pass on details. Final copy tweaks, color refinements, layout micro-adjustments.
If you're on round 7, you've usually been making rewriting mistakes instead of editing mistakes. Stop, look at what's actually wrong, fix only that.
The "keep, fix, replace" pattern
For every Claude design output, before requesting changes, categorize what you see into three buckets:
Keep
Specific elements that work. Name them explicitly. "The hero headline is exactly right." "The three-column feature layout works." This protects them from being inadvertently rewritten.
Fix
Elements that are 80% there but need adjustment. Name them and describe the fix. "The second CTA is too long — cut to under 5 words." "The third feature column's copy is generic — rewrite specifically about the API."
Replace
Elements that need to be redone from scratch. Name them and describe the new direction. "The pricing section doesn't work — replace with a single-tier focus, no comparison table, just one big price and one CTA."
The redirect prompt template
KEEP exactly as is: - [Element 1] - [Element 2] FIX (keep structure, adjust): - [Element]: [specific change] - [Element]: [specific change] REPLACE (start over for this part): - [Element]: [new direction] Everything else stays the same.
The "everything else stays the same" line
Claude is great at producing variation. Sometimes too great — it'll regenerate things you liked because they're adjacent to the things you're fixing. The phrase "everything else stays the same" tells Claude to preserve what worked.
The iteration mindset
Design work with Claude is editing, not commissioning. You're not asking for a new version of the work — you're directing surgical updates to the version you have. Treating it this way is the single biggest unlock for going faster.
Working with real designers (when you need to).
Claude doesn't replace designers. It changes what you should hire designers for. The new division of labor saves you money, lets designers do better work, and produces better results.
The new division of labor
What you do with Claude
- Strategic decisions about what the design needs to do
- Section-by-section content and structure
- Brand direction (colors, type, voice)
- First-draft layouts and mockups
- Variations and exploration
What you hire a designer for
- Original creative direction that distinguishes you
- Custom illustration and imagery
- The final 20% polish that separates "presentable" from "exceptional"
- Design systems that need to scale across many touchpoints
- Print production and execution complexity
How to brief a designer when you've done Claude work first
The brief becomes radically tighter. You're not asking the designer to figure out what to do — you're asking them to execute work where 80% of the decisions are already made. This produces three benefits:
- Cheaper. Designers charge for strategic thinking. Less strategy needed = lower hours = lower cost. Often 30–50% reduction.
- Faster. No discovery phase. You're handing them a working brief on day one.
- Better. Designers do better work when the strategic problem is clear. You're letting them focus on craft instead of strategy.
The handoff package
When briefing a designer with Claude pre-work, hand over:
- The brand sheet (from walkthrough 1)
- The structured outline (from walkthrough 3, or equivalent)
- Reference brands — the same ones you gave Claude
- The Claude-generated mockup as a "this is the direction, make it better" reference
- Your "never do" aesthetic rules — things you've already decided you don't want
A designer with this package can produce a finished design in days, not weeks. And the result will be more aligned to your actual needs than what they'd produce starting from a generic kickoff conversation.
When to skip the human entirely
For these, Claude alone (or Claude + Canva, or Claude + Code) is sufficient:
- Internal documents and presentations
- Small-scale landing pages and one-pagers
- Social ad creative for testing
- Email templates
- Quick mockups for stakeholder feedback
When to bring a human in
For these, hire a real designer (even with Claude pre-work):
- Your primary brand identity
- Anything that becomes your public face long-term
- Customer-facing product UI
- Pitch decks for high-stakes fundraising
- Anything that needs to win an award or get attention as design
The honest version
Claude makes design faster and cheaper for 80% of working-professional design work. For the 20% where design IS the work — your brand, your product UI, your moment-of-truth marketing — hire a human, but hand them a Claude-prepared brief. You'll get a better designer's time on a better-defined problem. Everyone wins, including the designer.
Design module — done.
You've finished the Design module — and with it, the entire Add-On Pack.
What you walk away with
The honest map of capability.
Claude is a senior production designer with 10 years of experience. Great for 80% of working-professional design. Not where you go for original brand identity or design-led product work.
The design RCTC.
Same framework as prompting. Role, Context, Task, Constraints — adapted to visual outcomes. Always include a reference brand whose aesthetic you'd borrow.
Three walkthroughs.
Brand sheet from prose. Board slide deck from a memo. Landing-page mockup that designers can execute. Three patterns that cover most working-professional design work.
The new division of labor.
You do strategy with Claude. Designers do craft and execution. Briefs are tighter, costs are lower, results are better. Everyone wins.