Claude Design: Full Walkthrough. I'm blown away.
Greg Eisenberg conducts a live walkthrough of Claude Design, testing its wireframing, visual design, pitch deck, and video creation capabilities using a senior brain-training app idea called 'Senior Brains.' He finds the wireframing and deck generation exceptional, the visual designs strong, but the video capabilities underwhelming compared to competitors.
Summary
Greg Eisenberg hosts a live stream exploring Claude Design (claude.ai/design) in real time, using an app idea sourced from IdeaBrowser.com — a gamified brain-training app for seniors called 'Senior Brains.' He intentionally starts with a low-fidelity wireframe rather than jumping straight to high-fidelity designs, citing token efficiency and the value of constraints in product thinking.
The wireframe generation process begins with an impressive AI-driven questionnaire that asks targeted product questions about device type, gamification elements, accessibility features, mascot tone, and exercise types. Greg is visibly impressed by the quality of these questions, comparing them to what a skilled product manager would ask. Claude Design produces three distinct wireframe directions — Warm Stack (Direction A), Mascot Forward (Direction B), and Calendar Ritual First (Direction C) — mirroring an agency-style creative presentation. Audience voting in the live chat selects Direction A to proceed.
Greg then attempts to generate a high-fidelity version of Direction A, encountering several errors and bugs along the way. He documents these failures openly, contrasting his experience with polished YouTube tutorials that edit out such moments. He discovers that running two projects simultaneously causes Claude Design to stall or lose context, advising viewers to focus on one task at a time.
While waiting for the hi-fi designs to render, Greg creates a parallel project — a VC pitch deck targeting Sequoia Capital for a $2M seed raise. The resulting deck impresses him enormously, featuring market sizing, competitive landscape, go-to-market strategy (Meta ads targeting adult children), unit economics (LTV/CAC of 3.7x, 3.3-month payback), and a compelling narrative. He calls it 'probably the best deck I have ever seen created by any LLM.'
The hi-fi visual designs for Senior Brains eventually render and are well-received — featuring a mascot named Bean, large tap targets, family social features, and gamification elements. Greg uses the freehand annotation/drawing tool to request a 'Share this Win on Facebook' button, which Claude Design implements with strong copy.
Finally, Greg tests Claude Design's video ad creation capabilities, requesting a 30-second commercial for Senior Brains targeting adult children aged 35–45. The output is described as underwhelming and not cinematic enough for a TV or Meta ad. He compares it unfavorably to a competitor tool (eversince.ai), concluding that Claude Design is not suited for video production.
His final verdict: Claude Design is best-in-class for wireframing and pitch decks, very strong for visual app design, but rates only about 5/10 for video. He plans to continue using it and recommends others get hands-on experience with it.
Key Insights
- Greg argues that starting with low-fidelity wireframes before hi-fi designs is better practice because it saves tokens and helps establish product constraints — he explicitly says 'mid-fi wireframes are bad' and that you should go lo-fi or hi-fi, not in between.
- Greg is blown away by Claude Design's onboarding questionnaire, comparing it to what a skilled product manager would ask — covering device type, mascot tone, gamification elements, accessibility needs, and family caregiver prominence.
- Claude Design generates three distinct wireframe directions (Warm Stack, Mascot Forward, Calendar Ritual First) in an agency-style creative presentation format, which Greg says mirrors what a real design agency would deliver to a client.
- Greg discovers that running two Claude Design projects simultaneously causes one to stall and lose context, concluding that the tool does not support parallel task execution and users must focus on one task at a time.
- The AI-generated VC pitch deck for Senior Brains impresses Greg so much he calls it 'probably the best deck I have ever seen created by any LLM,' noting it included unit economics (LTV/CAC 3.7x), go-to-market strategy via Meta ads, and a competitive landscape analysis.
- Greg observes that Claude Design's questionnaire process is one of its most valuable features because it forces the user/product manager to clarify their thinking — when the questionnaire failed to reappear after a session switch, he flagged it as a meaningful product miss.
- The hi-fi visual designs for Senior Brains feature a mascot named Bean, large tap targets, family social cheering features, and gamification elements — Greg says these exceeded his expectations and rated about 90% of the way to production quality.
- Greg uses Claude Design's freehand annotation tool to request a 'Share this Win on Facebook' button, and notes that the AI's copy choice ('Share this Win' vs. 'Share on Facebook') shows conversion-rate awareness consistent with social product best practices.
- Greg rates Claude Design's video ad creation at approximately 5 out of 10, saying it is not cinematic enough for TV or Meta ads, and directly compares it unfavorably to eversince.ai which produces hyperrealistic video output.
- Greg notes that the live stream format reveals real bugs and limitations that polished YouTube tutorials edit out — errors occurred multiple times during the session including rendering failures and context loss when switching between projects.
- Greg argues that the senior adult demographic (65+) is massively underserved by software, and sees Senior Brains as a potential $5–15M ARR business because the adult child is the buyer while the senior is the user — a gifting dynamic well-suited to Meta advertising.
- Greg observes that Claude Design's seamless navigation breaks down when leaving and re-entering a project folder — it loses conversation context and feels disconnected, which he explicitly flags as a UX miss for the product.
Topics
Full transcript available for MurmurCast members
Sign Up to Access