How the product designer role shifts when everyone is a professional vibe coder.

AI was taking the world by storm, and after a period of gentle little nudges from the top to use it, came a shove into the deep end.

At one of the All Hands, our CEO excitedly confessed that he spent over a hundred hours vibe coding, gushing about how it was the most fun he's had since starting the company. It was a clear signal. A few days later, the Product Org was informed that we all have licenses to Cursor and Claude Code now, and multiple AI learning meetings started popping up on everyone's calendars. It felt like it was only a matter of time before AI fluency becomes a part of our performance reviews. We were strongly urged to experiment and work these tools into our processes, but without much guidance from Leadership.

People were anxious. For the designers whose careers were centered around the PED (Product, Engineering, Design) model, they were getting worried about the blurred lines between functions. There were already situations where product managers were creating prototypes in silo and only handing them off to designers to "clean them up". Are we going back to the world where designers are seen as pixel pushers? If product managers can vibe code prototypes and product designers can create PRDs, who owns what these days?

The tools don't separate the functions. The outcomes do.

I think we're starting to see some form of an identity crisis. When anyone can spin up a plausible layout, draft copy, vibe code a prototype, or even push code to production, the traditional PED model starts to feel brittle. Once boundaries are blurred, ownership does too.

But the concern assumes that roles are defined by tools and activities. They shouldn't (or at least I don't think so). Because tools and activities are the first to get reshaped by automation, I think a more durable way to frame it is that each role owns an outcome. It's a belief that I've held onto for years.

Product Managers own the why.

The core responsibility of product is not writing PRDs, running standups, or even prioritizing the backlog because these are things that AI can easily do. The real responsibility is defining why something should exist at all. Product managers advocate why this problem matters, why it's worth solving now, why this approach over alternatives, and why it will create value for users and the business.

In an AI-accelerated world, the volume of possible ideas explodes. The space gets saturated with features and working prototypes, which makes it even more important to pare down on the ideas that matter. Someone must cut through noise, connect user needs to business strategy, and make coherent bets under uncertainty. AI can generate options, but it can't be accountable for conviction, a word that I've been using a lot more in the past few years. That's what I expect out of product managers.

Designers own the what.

If Product defines why something should exist, Design defines what it actually is and does. Design shapes the the experience, the interaction model, the system of decisions that make a product usable, coherent, and meaningful.

AI can generate interfaces, but the responsibility of design is not to produce artifacts; it's to ensure that what gets built expresses the intent. What should the user feel? What mental model should this create? What tradeoffs are we making in simplicity vs. power? All of these get considered and end up manifesting in the final product.

As AI increases the speed of exploration, designers become even more critical as curators. They turn an infinite design space into a deliberate, human-centered outcome. This is an area designers tune into more because it's what we signed up for. We pay close attention to the visuals, the hierarchy, and the interactions because we've got the taste and have trained for this.

Engineers own the how.

I think Engineering's role has gotten even more important as well. “Writing code” is a shallow interpretation of their job, which makes it seem especially vulnerable in the age of AI-assisted development. But code is just the medium. The real responsibility is determining how something works—how it scales, how it performs, how it remains reliable, secure, and adaptable over time, especially when more folks are touching the code. Sure, AI can generate code and make implementation faster, but it doesn't own up to the consequences of those decisions. Engineers do.

This elevates the importance of engineering judgement. The “how” becomes more about building systems with the right guardrails and making them last.

Where Design Starts to Feel Different

It's hard to say whether or not AI is having a big net positive and tangible effect on the product development process at my company right now. I'm experimenting like everyone else, but have found some things sticking for me. I'm doing less of the stuff that feels mechanical and having all of that automated. Things like taking notes for meetings, organizing insights, and creating research decks are just some of the items I can offload with tools like Granola and NotebookLM.

The big unlock in all of this, is speed to execution. Design has always been about de-risking and screening out the bad ideas. Now, we can do that even faster and earlier in the process. I can create prototypes (without ever wireframing in Figma) that feel like the real deal and put them in front of users for testing quickly.

It's important to be wary of echo chambers. If AI was a person, it would be a people pleaser, giving toxic positivity sometimes. I think healthy opposition creates growth and opportunities, so I always explore alternatives despite what AI spits out. I will use Figma Make to help me visualize different variants without losing intent. Honestly, I'm probably in the same boat as other designers, where we're all driven by curiosity and corporate pressure, while still figuring out how to codify some of these workflows.

Designers still need to think about intent and polish.

In my experience, AI isn't that great at picking up intent and edge cases. That part still needs human involvement. We need to apply judgement, create direction, and communicate meaning around the output of our work, so the gap with AI is still pretty real.

What I am finding a bit more freeing is spending more time in polish in the best sense—not ornament, but the signal that the product respects a user's time and energy. I'm obsessed with introducing delightfulness at every moment in the experience wherever I can. I think AI-native workflows speed us up towards obvious iterations so designers can spend more cycles on making small details that add up, on consistency across flows, and making the default path feel unmissable. The opportunity is finally having bandwidth for both purpose and polish.

The bar for Design is now higher.

As of today, I don't think AI is ready to completely replace high functioning design teams. If anything, AI forces each function to operate at a higher level of abstraction (and faster too).

I think Design still has business in this world. The tools are getting more powerful, which means our expectation of judgement needs to be as well. Do these AI solutions play nicely within the bigger picture of how we want our product to be? Is it taking into account the holistic experience across other surfaces? A huge part of our role is social. We still need to align our teammates and get them rallying in a direction. We still need to advocate for users in business conversations. We still need to influence decisions without having authority sometimes. Designers are the ones sitting in the room negotiating priorities and building trust, not the tools. In other words, not AI.