Enough. It\u2019s time to blow up the Dev-Star and roll a Snowball<\/strong> instead.<\/p>\n\n\n\n Why does the old hand-off model fail so spectacularly now? Because AI changed the rules<\/strong>:<\/p>\n\n\n\n Old-world relay race: research \u2192 design \u2192 PM \u2192 dev was fine when you were building deterministic features. In AI-first development, it\u2019s a disaster. Nuance gets lost, assumptions drift, and by the time dev ships, you\u2019ve got something no one recognizes.<\/p>\n\n\n\n It\u2019s exactly like a game of \u201cTelephone Pictionary\u201d \u2014 what starts as \u201cHis heart stopped when she came into the room\u201d<\/em> somehow mutates into \u201cVideo killed the radio star.\u201d<\/em><\/p>\n\n\n\n That\u2019s the hand-off problem in one meme. Each step adds distortion until the final output is convenient but wrong<\/strong>.<\/p>\n\n\n\n Hilarious for onlookers.<\/p>\n\n\n\n Not nearly so funny for your company (and your career).<\/p>\n\n\n\n (And that\u2019s why this post is called \u201cSnowball Killed the Dev-Star.\u201d<\/strong> Just like Video Killed the Radio Star,<\/em> it\u2019s my tongue-in-cheek way of saying: the old era is gone. The hand-off-heavy Dev-Star model is obsolete. The only way forward is rolling the Snowball together.)<\/p>\n\n\n\n Here\u2019s what happens when you cling to Figma-first thinking: users are asked to evaluate surface decoration<\/strong> while the actual AI behavior<\/em> stays hidden.<\/p>\n\n\n\n <\/p>\n\u201cHere are two AI Solutions, A and B. Which do you like better?\u201d<\/em><\/cite><\/blockquote>\n\n\n\n That\u2019s what wireframe testing has become. Rain spouts design. Dragon or Gargoyle? While your AI plumbing empties into the elevator shaft<\/strong>.<\/p>\n\n\n\n Sure, you\u2019ll get lots of opinions about which rain spout looks cooler. But the heart of the product<\/strong> \u2014 the AI\u2019s decisions, the data flows, the trust dynamics \u2014 remains a black box. Users can\u2019t meaningfully respond to that, so we confuse shallow preferences with deep validation.<\/p>\n\n\n\n It reduces UX research to a macaroni-and-glitter project: \u201cDo you like it, Mom? You like it, right?\u201d<\/em><\/p>\n\n\n\n This is unacceptable.<\/strong><\/p>\n\n\n\n The fix is obvious: get customers interacting with live AI outputs<\/em> \u2014 vibe-coded prototypes, Snowball iterations, real data-driven behaviors. That\u2019s how you surface insights that actually matter.<\/p>\n\n\n\n Developers siloed in their Dev-Star aren\u2019t the heroes anymore.<\/p>\n\n\n\n <\/p>\n\u201cIn AI-first products, the design is how it works.<\/strong> UX, Dev, and Data are inseparable. If you\u2019re not collaborating across roles, you\u2019re flying blind.\u201d<\/em><\/cite><\/blockquote>\n\n\n\n AI projects fail because devs in the silo of their Dev-Star optimize for model accuracy while ignoring user value, while PMs and designers (from outside the Dev-Star) hand off pretty UIs that don\u2019t reflect AI uncertainty. It\u2019s a vicious cycle \u2014 one that 85% of AI projects never escape.<\/p>\n\n\n\n <\/p>\n\u201cHow AI works\u201d is no longer just dev responsibility. AI is simply too important to be left to Developers and Data Scientists. <\/strong>Everyone on the team needs to be intimately familiar with \u201chow it works\u201d and develop the skills to understand enough context to ask the right questions.\u201d<\/em><\/cite><\/blockquote>\n\n\n\n The issues with the traditional 3-in-a-box development process have always been there; AI just exposed them, so they are now impossible to miss. \u201c3-in-a-box\u201d reduces the probability you will get your AI project \u201cright\u201d to the chances of catching a black cat in a dark room.<\/p>\n\n\n\n Blindfolded.<\/p>\n\n\n\n While wearing welding gloves.<\/p>\n\n\n\n It\u2019s no surprise most AI projects faceplant \u2014 multiple studies show 85%+ of AI initiatives never deliver ROI<\/strong>. (Honestly, I\u2019m surprised any succeed!)<\/p>\n\n\n\n Expensive lesson:<\/strong> just look at some high-profile AI flameouts:<\/p>\n\n\n\n Each disaster had the same root cause<\/strong>: teams applied old-school, siloed thinking to AI problems and got old-school failure<\/em> as a result. AI projects implode when built in a vacuum.<\/p>\n\n\n\n So how do we avoid Dev-Star hell? By embracing a fundamentally different way of working \u2014 the Snowball model<\/strong>. Instead of big hand-offs and isolated silos, the Snowball model means building together, iterating constantly, and involving real users from day one<\/strong>.<\/p>\n\n\n\n The team (PM, UX, Dev, Data Science \u2014 everyone<\/strong>) works as one unit, folding successive layers<\/strong> of insight and functionality into the product like packing a snowball, while customers call BS in real time<\/em>. The product keeps moving forward in small, testable increments, growing richer and more robust with each \u201croll\u201d of the Snowball. No big bang, no big flop \u2014 just continuous learning and adjusting.<\/p>\n\n\n\n This isn\u2019t touchy-feely talk; it\u2019s a call to arms for UX and product people<\/strong>. AI isn\u2019t the death of design or PM \u2014 it\u2019s their rebirth as a team sport<\/strong>. To thrive, we all have to get our hands dirty together: understanding data, designing with AI behavior in mind, and yes, even coding<\/strong> (more on that below). In an AI-first world, \u201cdesign is how it works\u201d<\/strong> \u2014 the UX is inseparable from the code and data. If you\u2019re not in the code, if your data scientists and designers and PMs aren\u2019t in constant collaboration, you\u2019re flying blind.<\/p>\n\n\n\n Ready to ditch the hand-off? Here are the Snowball principles that will save your AI project (and your sanity):<\/p>\n\n\n\n Notice what these principles have in common: they\u2019re all about action, collaboration, and reality checks<\/strong>. You\u2019re constantly building and<\/em> validating, together with your team and your users. That\u2019s the Snowball method in a nutshell.<\/p>\n\n\n\n One of the most radical (and freeing) ideas of the Snowball model is this: your first prototype should have zero traditional UI<\/strong>. Instead of opening Figma, you open a text file. Instead of wireframing, you’re uploading CSVs to ChatGPT and creating RAG files.<\/p>\n\n\n\n Starting with LLM + RAGs isn’t the death of design \u2014 it’s design evolution at its purest form<\/strong>. While traditional designers debate button placement, evolved designers are validating entire product concepts armed with a deep understanding of the bespoke data that provides the core of the competitive advantage for your company in the AI-First universe.<\/p>\n\n\n\n One design leader I know tested an AI customer service solution using only Claude and their existing FAQ database \u2014 500 users eagerly validated the concept before a single mockup existed. Another UXer used GPT-4 plus their company\u2019s insurance product catalog to prototype a shopping assistant that generated multiple pre-orders with zero custom UI.<\/p>\n\n\n\n This approach embodies the new Snowball model perfectly: start with the small, dense core (your data), add LLM capabilities with the existing chat Interface, and toss the snowball at some customers to see what sticks. <\/p>\n\n\n\n The fastest way to test an AI product concept is by simulating the AI itself<\/em>.<\/p>\n\n\n\n By the time you consider designing a UI, you\u2019ve already validated the core UX<\/strong>: product-market fit, the overall flow, the AI\u2019s output, the tone, and what users actually want from your solution. You can adjust prompts or data in real-time when talking to a user (\u201cOh, it\u2019s too verbose? Let me tweak that now… how about this?\u2026 What do you think of those headers?\u201d<\/em>) \u2014 try doing that with a static prototype!<\/p>\n\n\n\n In a matter of days, you\u2019ll have nailed down things that truly matter (e.g., what the AI should say and do, what \u201cgood\u201d looks like to the user, etc.) while competitors are still arguing over button colors and playing Telephone Pictionary with requirements. The designers who embrace this data-first AI-first prototyping<\/strong> aren\u2019t abandoning their craft \u2014 they\u2019re elevating it<\/em>. They focus on what really<\/em> matters to UX (the heart of the<\/strong> experience<\/strong>, not the gargoyle rain spouts) and validate real solutions<\/strong> instead of churning out pretty placeholders choked with lorem-ipsum<\/em>.<\/p>\n\n\n\n So ask yourself: Will you evolve with the discipline, or cling to the sinking Figma Titanic while others ship working AI-driven products?<\/strong> Because make no mistake, the iceberg has already hit. The old UI-centric design ship is taking on water.<\/p>\n\n\n\n This is a rallying cry to my fellow UX professionals. AI is flipping our field upside down. If you\u2019re the \u201cUX designer\u201d who only makes static Figma slideshows, you\u2019re about to go the way of the telegraph operator in the smartphone era<\/strong>. <\/p>\n\n\n\n Harsh but true.<\/p>\n\n\n\n In the Snowball world, the designers who learn to code (\u201cvibe coding\u201d with AI help \u2014 see below) and embrace low-fi experimentation are running circles around those stuck in old workflows<\/strong>. <\/p>\n\n\n\n They\u2019re shipping AI products 10\u00d7 faster<\/strong> than their peers, because they\u2019re not waiting on hand-offs and Telephone Pictionary that entails \u2014 they\u2019re helping their teams to create functional prototypes, getting them in front of customers, and iterating in hours, not weeks.<\/p>\n\n\n\n Let\u2019s be clear: this isn\u2019t about turning UXers into full-time software engineers. It\u2019s about breaking down the Dev-Star silos that say \u201cdesigners design, engineers build.\u201d Modern UXers can and should dip into implementation<\/strong>, especially now that tools like Claude and GPT-4 can help generate high-quality working code. We call it \u201cvibe coding\u201d<\/strong> \u2014 designers translating their vision directly into working prototypes, using AI to fill in the gaps. When you do this, you transform from a mere wireframe artist<\/em> into a product catalyst<\/strong>. While others are still redlining specs or begging devs to review their mocks, you\u2019re already testing a live prototype with real users. And guess what? The market rewards it.<\/strong> Designers who can straddle both worlds command higher salaries and bigger roles<\/strong>, because they drive innovation instead of decorating it. Those who cling to just pushing pixels are, frankly, being automated away<\/em> by the very AI tools they ignored.<\/p>\n\n\n\n I know this sounds intense. It is<\/em> intense! But it\u2019s also exciting<\/em>. This is a chance to reinvent what it means to be a UX designer or product person. The ones who seize this moment are going to define the next generation of products. The ones who don\u2019t\u2026 well, you probably won\u2019t hear from them in a few years.<\/p>\n\n\n\n The AI revolution isn\u2019t coming \u2014 it\u2019s here<\/strong>. The only question now is whether you<\/em> will be one of those who shape it. The old way is dying (good riddance); a new, more dynamic, more empowering way of working is emerging. If you\u2019re ready to grab that mantle, I\u2019m here to help you make it happen. Don\u2019t let your career sink with the old ship. Let\u2019s build something incredible<\/strong> instead.<\/p>\n\n\n\n Code talks. Bullshit walks.<\/strong> The future belongs to those who build<\/strong>.<\/p>\n\n\n\n Are you in? The article originally appeared on UX for AI<\/a>.<\/em><\/p>\n\n\n\n![\"\"](\"https:\/\/uxmag.com\/wp-content\/uploads\/2025\/09\/Image-1-1-1024x537.png\")
\n
Hand-off \u201ctelephone pictionary\u201d kills momentum<\/h2>\n\n\n\n
![\"\"](\"https:\/\/uxmag.com\/wp-content\/uploads\/2025\/09\/Image-2-1024x711.png\")
The worst part: customers are reduced to choosing rain spouts<\/h2>\n\n\n\n
\n
![\"\"](\"https:\/\/uxmag.com\/wp-content\/uploads\/2025\/09\/Image-3-1-1024x381.png\")
\u201cDev-star\u201d coding isolation is a dangerous myth<\/h2>\n\n\n\n
\n
\n
\n
Roll the snowball: a new way to build AI products<\/h2>\n\n\n\n
![\"\"](\"https:\/\/uxmag.com\/wp-content\/uploads\/2025\/09\/Image-4-1024x711.png\")
The snowball manifesto: 10 principles for AI-first teams<\/h2>\n\n\n\n
\n
Start with the heart (not the rain spouts)<\/h2>\n\n\n\n
This is Wizard of Oz on steroids<\/h2>\n\n\n\n
UX designers: code or die (evolve or fade out)<\/h2>\n\n\n\n
Ready to lead the AI-first UX revolution?<\/h2>\n\n\n\n
Greg<\/p>\n\n\n\n