One of the clearest ways AI is already changing how we work as designers is in the rapid acceleration of prototyping. That’s not an abstract statement, it’s something I’ve experienced firsthand in a high-pressure B2B SaaS environment, where speed and clarity often make the difference between stakeholder buy-in and endless looping.
In particular, tools like Lovable have fundamentally shifted how I think about turning early ideas into something testable. These are generative prototyping platforms that can spin up surprisingly realistic UI flows based on screenshots, style prompts, or basic user journeys. You feed it some structure, and it gives you a clickable, styled prototype, often within minutes.
That doesn’t just save time, it changes the type of conversations you can have early in the process.
A Real Example: From Pitch to Prototype in a Day
A few months ago, I was helping prepare for a pitch with a potential NHS client. The sales team wanted to show how our product might evolve to support a newly emerging staffing model. In the past, I’d have mocked something up in Figma, stitched together a few screens, and wrapped it in a static PDF. It would’ve looked nice, but it wouldn’t have told the full story.
This time, I used Lovable.
Starting with a few screenshots of our existing product and a description of the proposed flow, I generated a full, interactive prototype by the end of the day. It wasn’t pixel-perfect, but it was real enough to walk them through the experience. That speed allowed us to have a much richer conversation, focused on the idea, not the artefact.
The Upsides
Here’s where these tools really shine:
Fast ideation: You can test multiple directions without getting too attached.
Sales enablement: Prototypes help make ideas tangible for non-technical stakeholders.
Cross-functional empowerment: Product managers, strategy leads, even ops teams. can prototype ideas themselves, as long as guardrails and systems are in place.
Greenfield exploration: Without requiring roadmap investment or engineering resourcing, you can model POCs and test-market directions quickly.
The low barrier to entry means more people in the business can contribute to early thinking, without clogging up design or engineering capacity. That’s a huge unlock when you’re moving fast.
But Don’t Be Fooled by the Polish
The problem is: it looks too good.
I’ve seen clients and internal teams mistake these prototypes for near-production. Because the UI is polished and interactive, it creates a false sense of completeness. Suddenly, you're having to manage expectations you didn’t set, or worse, defend timelines that were never realistic.
Even internally, these tools can lead to people skipping critical steps, feasibility reviews, tech estimates, or user validation. And while tools like Lovable offer code export and Git integration, let’s be honest: that code is rarely usable. It’s often bloated, poorly structured, or dependent on Tailwind-based visuals that lack brand fidelity and accessibility.
I’ve worked with multiple engineering teams who’ve tried to bridge the gap from Lovable to production. So far, none have found a clean way to do it without major rework or reverse engineering. In some cases, it’s faster to start from scratch.
My Mantra to use It Effectively
I still use Lovable regularly. But I use it with clear boundaries. Here’s how I keep it productive:
Set expectations early. Every time I share a Lovable prototype, I clarify: “This is for direction only. It’s not production-ready.”
Use it to start conversations, not end them. These prototypes are meant to provoke discussion — not make final decisions.
Pair with engineering input. If it’s something we might ship, I involve an engineer early — even if it’s just to say, “Is this even feasible?”
Expect iteration. It usually takes 2–3 refinement cycles to get something that fits your product’s tone and style.
Final Takeaway
Generative prototyping has become one of the most powerful tools in my kit. It’s changed how quickly I can test, sell, and validate ideas. But it’s not a replacement for process, and it’s definitely not a shortcut to delivery.
Used thoughtfully, it can unlock creativity and collaboration across your org. Used blindly, it creates risk, rework, and confusion.
AI can help you move fast. But moving fast in the wrong direction is still just… going the wrong way.