Wireframing Tools For UX

I've been experimenting with Claude Code as a partner for rapid design exploration. Loaded with the essential project context, I ask it to explore a problem and propose five alternative solutions as sketch wireframes. This helps me think through the solution space faster and hone in on a promising direction. Staying at wireframe fidelity also keeps me from getting hung up on the details Claude Code still stumbles on. A real risk with designing with AI agents is getting too fixated on an early direction and polishing the wrong thing. Once “we” find a solid direction, I work with Claude Code to polish it with the right visual context, whether that's a design system or UI references. Often I'll also open a new Figma file and take one or two of the directions further there. I never take a direction from Claude Code as given; I always add my own thinking to it. Sketchy wireframes act as a bridge: from using an AI agent as a research assistant to exploring solutions with that same context still loaded. It feels like a way of using Claude Code as a real thinking partner I guide — not as a hands-off, faulty UI generator. The inspiration for this came from Jenny Wen, design lead for Claude at Anthropic, who described a similar process on Lenny's Podcast recently.

Some of you disagreed with my last post. Fair. Let's talk. Let me explain the topic a bit more and give you a deep dive into how I see the new process. The old way: Think → Research → Wireframe → Design → Spec → Hand off → Build → Test → Iterate Weeks. Sometimes months. Before anyone touches real code. The new way: 👉 Step 1: Start with a problem, not a doc. I don't need a full PRD. I need one thing. Example: "𝘗𝘦𝘰𝘱𝘭𝘦 𝘴𝘵𝘳𝘶𝘨𝘨𝘭𝘦 𝘵𝘰 𝘨𝘦𝘵 𝘩𝘰𝘯𝘦𝘴𝘵 𝘧𝘦𝘦𝘥𝘣𝘢𝘤𝘬 𝘰𝘯 𝘵𝘩𝘦𝘪𝘳 𝘱𝘰𝘳𝘵𝘧𝘰𝘭𝘪𝘰." That's it. That's the brief. 👉 Step 2: Build the ugliest working version. I open Lovable or Cursor and prompt my way to a prototype. Not a mockup. Not a Figma file. A real, clickable, functional thing. 30 minutes. Maybe an hour. 👉 Step 3: Use it. Don't refine it. Don't show it to anyone yet. Use it yourself like a real user would. Click every button. Try to break it. Feel where it's awkward. 👉 Step 4: Now design. This is where design skill actually matters. You're not guessing what the experience should feel like. You already know because you felt it. Now you fix what's broken, remove what's unnecessary, and polish what works. Maybe pivot or try other solutions. 👉 Step 5: Show it, don't spec it. Instead of a 20-page spec, I send a link. "Here, try this. What's confusing?" Real feedback on a real thing beats hypothetical feedback on a hypothetical thing every single time. 👉 Step 6: Iterate in minutes, not weeks. Here's where this workflow really pulls ahead. Someone says, "This flow is confusing." You don't update a Figma file, write a ticket, and wait for the next sprint. You open Cursor, fix it, and send a new link. Same conversation. Same day. The feedback loop goes from weeks to hours. Sometimes minutes. And each round gets sharper because you're iterating on something real. 3-4 rounds of this, and you have something more validated than most products get after months of traditional process. 👉 Step 7: Document what you built, not what you plan to build. Documentation becomes a record, not a prediction. It's accurate because the thing already exists. You can do it at the end or during the process. Why this works: You make decisions with information instead of assumptions. You eliminate 80% of the back-and-forth. You design from experience, not imagination. And you iterate at the speed of conversation, not the speed of sprints. Why it feels wrong at first: Because we were trained to think before we build. And thinking first felt responsible. But we did that because we couldn't build. Now we can. And I don't think it's about ignoring thinking. (𝘔𝘢𝘯𝘺 𝘰𝘧 𝘺𝘰𝘶 𝘢𝘤𝘤𝘶𝘴𝘦𝘥 𝘮𝘦 𝘰𝘧 𝘵𝘩𝘢𝘵) I believe it's about doing it at every step. Refining it based on real feedback. Insights you can get internally and from user testing. If you're still reading this, let me know what you think about it all. ✌️

𝗔𝗜 𝗷𝘂𝘀𝘁 𝗱𝗲𝘀𝗶𝗴𝗻𝗲𝗱 𝗮𝗻 𝗮𝗽𝗽 𝗳𝗼𝗿 𝗺𝗲 𝗶𝗻 𝟱 𝗺𝗶𝗻𝘂𝘁𝗲𝘀. 𝗛𝗲𝗿𝗲’𝘀 𝘄𝗵𝗮𝘁 𝗵𝗮𝗽𝗽𝗲𝗻𝗲𝗱... Last week, we were all talking about the 𝘤𝘰𝘮𝘦𝘣𝘢𝘤𝘬 𝘰𝘧 𝘜𝘐 — from Material 3’s brutalist influence to Airbnb’s nostalgic skeuomorphic icons. And then, yesterday, Google dropped something in a completely different direction: 𝗦𝘁𝗶𝘁𝗰𝗵 — an AI tool that builds user interfaces from simple prompts. I gave it a 5-minute test. Prompted: “An app to keep track of my plants.” It generated 5 basic but functional screens I could edit right away. Was it perfect? No. Fonts, colors, and layout options are still limited — it feels more like a fancy wireframe than a polished UI. But that’s exactly what makes it 𝗶𝗻𝗰𝗿𝗲𝗱𝗶𝗯𝗹𝘆 𝘂𝘀𝗲𝗳𝘂𝗹 𝗳𝗼𝗿 𝗲𝗮𝗿𝗹𝘆-𝘀𝘁𝗮𝗴𝗲 𝗽𝗿𝗼𝘁𝗼𝘁𝘆𝗽𝗶𝗻𝗴. I can see tools like Stitch being a great companion during Design Sprints or ideation workshops — quickly turning ideas into something visual we can test and iterate on. It’s just the beginning, but it’s clear: AI is rapidly transforming how we design. And as designers, we need to evolve too — from pushing pixels to 𝗽𝗿𝗼𝗺𝗽𝘁𝗶𝗻𝗴 𝗶𝗻𝘁𝗲𝗻𝘁𝗶𝗼𝗻𝗮𝗹𝗹𝘆, 𝗰𝗿𝗮𝗳𝘁𝗶𝗻𝗴 𝗲𝘅𝗽𝗲𝗿𝗶𝗲𝗻𝗰𝗲𝘀, and 𝗳𝗮𝗰𝗶𝗹𝗶𝘁𝗮𝘁𝗶𝗻𝗴 𝗱𝗲𝗰𝗶𝘀𝗶𝗼𝗻𝘀. 𝗖𝘂𝗿𝗶𝗼𝘂𝘀 𝘁𝗼 𝗵𝗲𝗮𝗿 𝗳𝗿𝗼𝗺 𝗼𝘁𝗵𝗲𝗿𝘀: 𝗔𝗿𝗲 𝘁𝗼𝗼𝗹𝘀 𝗹𝗶𝗸𝗲 𝘁𝗵𝗶𝘀 𝗲𝘅𝗰𝗶𝘁𝗶𝗻𝗴, 𝗶𝗻𝘁𝗶𝗺𝗶𝗱𝗮𝘁𝗶𝗻𝗴, 𝗼𝗿 𝗯𝗼𝘁𝗵? Would you use it during a Sprint? #UXDesign #DesignSprint #RapidPrototyping #GoogleStitch #AItools #DesignLeadership #WorkshopFacilitation #StrategicDesign #PromptDesign #DesignTools #ProductDesign #UXUI #DesignWithAI

I spent the last 48 hours deep in Claude Design.  I don't need half my tools anymore. This is not an exaggeration. Anthropic dropped something significant on Thursday. Not a model update. Not a new feature buried in settings.  A full design environment - prototypes, decks, wireframes, one-pagers - powered by conversation. I went in skeptical.  I came out restructuring my entire workflow. Here is the 6-step process that actually works - thank you Charlie Hills 🦩: 1. Set up your Design System first. Claude reads your codebase and brand files during onboarding.  Every project after that inherits your colors, typography, and components automatically.  You set it up once.  It follows you everywhere. 2. Start in Wireframe. Then go High-fidelity.  Lock your layout fast, then duplicate into polish mode.  This alone saved me hours of rework I used to do constantly in Figma. 3. Upload your references.  Sketches, PDFs, a live URL, your codebase. It pulls everything in as context.  The web capture tool lets you grab elements straight from your actual product. 4. Write a specific brief.  This is still a prompting game. Vague gets you generic.  I wrote detailed briefs and got work that genuinely surprised me. 5. Refine live.  Comment on specific elements.  Edit text directly.  Use adjustment knobs Claude builds on the fly for spacing, color, layout. Apply a change to one element and push it across the full design. 6. Hand off to Claude Code.  When it is ready to build, one instruction packages everything into a bundle and passes it to Claude Code for production. The Figma loop I have used for years got compressed into a single conversation.  The Canva sessions I used to open just to make something presentable? I did not open it once. What genuinely shocked me was not the first output.  It was staying inside the revision cycle with it.  It does not just generate and leave. It thinks with you. The gap between having an idea and showing something real used to cost days.  It does not anymore. If you are on Pro, Max, Team or Enterprise, go try it now.  It is already inside claude.ai. Palette icon, left sidebar. Have you tried Claude Design yet? ♻️ Repost to your network 🔔 Follow Ranjana Sharma for practical takes on AI, leadership, and making this work in the real world

Vibe designing and how I'm building something new! For the last few months, I've been working on something huge with a team of 3. The flood of new AI tools has made it possible to accomplish in 3 months what would have taken 20+ people and 12+ months. Combining my 20+ years experience, I've 10X'd my influence over our product. Here's how I designed with AI: 𝗜𝗻𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻 𝗔𝗿𝗰𝗵𝗶𝘁𝗲𝗰𝘁 I explored concepts with Claude 3.7 Sonnet by: • Creating a project in Claude to maintain context • Starting with: "Make a mobile-friendly prototype..." + detailed requirements • Refining until I had comprehensive product specs + a clickable prototype Claude excels as an information architect and UX writer. Visual designer, not so much. 𝗩𝗶𝘀𝘂𝗮𝗹 𝗗𝗲𝘀𝗶𝗴𝗻𝗲𝗿 In Figma, I transformed Claude's wireframes into a product users would want. Maintaining prototypes became tedious as feedback rolled in. In conversations with Dixon Lo, Scott Lederer and Joann Wu discussing UX in 2025, I learned about many new tools like Relume.io and Bolt.new as ways to start “vibe designing”. 𝗨𝗫 𝗗𝗲𝘀𝗶𝗴𝗻 𝗘𝗻𝗴𝗶𝗻𝗲𝗲𝗿 Using Bolt, I created a functional React prototype: • Solved UX issues with simple prompts • Deployed to Netlify • Got it into users' hands quickly ⚠️ Bolt can get expensive ($200/month), but cheaper than hiring a developer. It's also a bit of a Monkey's Paw as you have to be careful what you ask for. It will perform a miracle and at the same time rip out 3 hours of work by accident. 𝗨𝗫 𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵𝗲𝗿 With a functional prototype, I: • Created test scripts using Claude • Used Otter.ai for observation notes • Extracted key insights • Incorporated changes before handing to our CTO Greg Karlin for development The speed was jaw-dropping compared to my big tech days. The difference is remarkable. 𝗧𝗵𝗲 𝗘𝘃𝗼𝗹𝘂𝘁𝗶𝗼𝗻 𝗼𝗳 𝗗𝗲𝘀𝗶𝗴𝗻 In 30 years, I've seen many shifts - from getting jobs just for knowing Photoshop in 1996, to the mobile boom, to robust design systems (I was Google's 4th visual designer working with Margaret Gould Stewart, Doug Bowman and others to create Googles first design guidelines.) Then AI entered the scene, and by 2023, many of us found ourselves spinning uncontrollably through space like Sandra Bullock in Gravity. The pace accelerated, transforming our approach overnight. I'm embracing the change. With experience and the right tools, we'll navigate this shift like the ones before. A year from now, everything will have transformed again—but evolution is what makes this field fascinating. While layoffs are concerning, it's exciting to create world-class products with limited resources. VCs can invest in more opportunities for less capital, creating an innovation explosion. What AI tools are you using? I'd love to hear your experiences! #AIinDesign #ProductDesign #UXDesign #GenerativeAI #vibedesign #vibeengineering *This post was AI assisted, not generated.

How to Use AI to Design UI Like a Boss These AI tools make your MVP look like it had a $50K design budget. Think you need a full UI/UX team to make your product look world class? Not anymore. AI is changing how founders and designers work, speeding up ideas, sharpening clarity, and turning rough concepts into real, testable screens in minutes. Here’s how to use it like a boss 👇 Wireframe in minutes Use Galileo AI/Stitch (usegalileo.ai) or Uizard (uizard.io). Describe your idea (“a dashboard for tracking SaaS metrics”) and watch it turn into a usable layout instantly. Iterate with Figma AI Drop your screens into Figma (figma.com) and use Magician to explore new layouts, refine spacing, and generate fresh design options fast. Write human microcopy Run your interface text through ChatGPT or Jasper (jasper.ai/tools) to make buttons, onboarding steps, and tooltips sound natural and on brand. Test usability instantly Upload your prototype to Maze (maze.co) or Useberry (useberry.com) and get immediate feedback on what’s working and what’s confusing. Automate accessibility checks Use Stark AI (getstark.co) to catch color contrast and readability issues early. Accessibility isn’t optional. It’s part of great design. Prototype like a pro Feed your screens into Framer (framer.com), Relume (relume.io), or Typedream (typedream.com) to create clickable prototypes you can pitch, demo, or validate with users. Because the faster you go from idea to interface, the faster you learn what people actually want. These tools won’t replace real design craft, but they’ll get you from zero to something real in hours instead of weeks. Follow me here on LinkedIn or subscribe to the Halo Newsletter for more founder tools, playbooks, and product strategies that give you an edge. #entrepreneurs #founders #b2bsaas #AI

If you still think building a startup prototype takes months, you’re probably not using the right tools. Last night, I went from zero to a working prototype in 30 minutes. No big team. No “stealth mode.” No months of planning. Just: 1️⃣ A real workflow bottleneck 2️⃣ AI-powered research 3️⃣ Instant wireframing & prototyping tools Here’s what actually happened (and how you can steal the playbook): Picked a real-world bottleneck: Instead of chasing “cool ideas,” I dug into the daily pain points of doctors in India—using ChatGPT & Gemini to synthesize thousands of words of research in minutes. Let AI play detective: Uploaded research docs to ChatGPT, got a prioritized list of the actual top bottlenecks. No assumptions, no wishful thinking. Prototyped instantly: Dropped the requirements into Lovable.dev, and in minutes had a clickable prototype ready to demo to real users. What did I learn? If you’re still spending weeks “validating” and “ideating,” you’re already behind. The right AI stack can take you from problem → insight → prototype before most people finish their deck. Ready to see how it works? Grab the full play-by-play PDF [Attached] Check the prototype link in the comments section. If you’re a founder still stuck on the whiteboard, try this workflow and break your own speed limits.

I just hit version 121 on a Figma Make. Here's what I learned about AI-assisted design. Most designers are still fighting Figma frames when they could be leveraging AI for rapid iteration. Here's the reality check I needed ↓ The Problem Everyone Faces: Designing complex user flows in Figma = death by a thousand frames • Duplicate a screen → make one change → update 47 other screens • Wire up interactions for every edge case • Spend hours on prototypes that test one simple concept Sound familiar? My AI Breakthrough Moment: I stopped asking AI to build everything at once. ❌ Before: "Create a comprehensive donation form with all features" ✅ Now: "Make this screen pixel-perfect. No functionality. Just nail the design." The difference? Night and day. What Actually Moves the Needle: 🎯 Start stupidly simple, then iterate 🎯 Use your design system properly (AI reads it better) 🎯 Focus on ONE thing you need to validate 🎯 Save comprehensive builds for later Plot Twist: AI doesn't replace design thinking. It amplifies it. While AI handles rapid concept generation, I focus on: → Understanding user needs → Making strategic decisions → Crafting experiences that solve real problems Question for you: What's your biggest frustration with current prototyping workflows? (Drop a comment - I read every single one and the conversation often sparks my best ideas) P.S. I wrote up the full process with examples and lessons learned. Link in comments 👇 #UXDesign #ProductDesign #AITools #DesignProcess #Prototyping