The 80% Problem: What Founders Don't Know About AI-Generated Apps

Forbes

08-04-2025

Josh Haas is the cofounder & co-CEO of Bubble, an AI visual dev platform that turns ideas into scalable apps—no coding required.
Imagine you're a founder with a big idea—a wellness app that matches users with custom nutrition plans based on health data. Once, this would have required a team of developers and $100,000 to build a prototype. Today? You type sentences into an AI tool, wait 30 seconds, and there it appears, seemingly ready to launch. It feels impossible, like the future has arrived ahead of schedule.
And in many ways, it has.
AI-powered tools have transformed what was once the domain of technical experts into a playground for anyone with an idea. As a founder with vision but no coding skills, you're suddenly free. No more being blocked by technical limitations. No more hunting for the perfect developer who might understand your vision. The barrier that kept countless ideas from ever seeing the light isn't just lowered; it's gone.
This moment not only feels transformative; it is transformative. My cofounder and I started Bubble, a 10-year-old platform that combines AI with visual, no-code development, because we knew that the future of software development lay in making computers speak our language, not the other way around. Flipping this script is the only way to make programming accessible to everyone and unlock economic opportunity for the masses, and AI has supercharged progress toward that vision.
There's no doubt that the future of software development is being built before our eyes. But there's a hidden truth that most AI app development platforms mask: What looks like a finished product is actually just the beginning.
One moment, you're cruising along, watching the AI make magic happen. But it's not long before that magic fades. That's when you find yourself spending hours refining prompts with increasingly specific language, only to get results that still miss the mark. The promise of "vibe coding" is shattered when the only way to fix the problem is to dive into the code—code that might as well be written in hieroglyphics.
This is the 80% problem.
That first impressive demo may make it look like 80% of the work is done, but it's often just 20% of what makes an app viable as a business. The rest—the part that creates actual user value and sustainability—is still ahead of you:
• A user experience that matches your vision but evolves in response to user feedback.
• Workflow logic that goes beyond surface-level functionality, like custom authentication flows or approval processes.
• Data management that supports real user volume, not sample data.
• Robust security features that protect your users and business.
This isn't just about adding features and changing colors. It's about transitioning from an app that demos well to an app that performs well in the real world.
The excitement of that first demo often masks the challenges ahead. You might not recognize that you've hit the 80% problem until you've invested significant time trying to push beyond it.
You'll know you're stuck when your app still bears all the visual hallmarks of AI generation, or it looks right but doesn't do everything you need. You spend an afternoon trying to add a simple dropdown menu that should take minutes. You find yourself frantically switching between different AI tools, hoping one will finally understand what you're asking for.
Growth becomes something you dread rather than celebrate because you're unsure if your app will crumble under real usage. User data protection? It wasn't built in. And that $20 worth of tokens you started with? You just burned through $500 more and still can't get that payment processing feature to work correctly. When something breaks, your only option is hiring a developer—exactly what you were trying to avoid.
In short: You're hitting the limits of what the platform can do, but you're nowhere near the limits of what your business needs to do.
The 80% problem isn't just a technical challenge—it will quash your business before it gets off the ground.
When progress stalls, you lose the most precious resources you have as a founder: time and momentum. What began as "free to start" becomes "expensive to finish" when you realize you need to hire specialists to overcome platform limitations.
Most critically, you miss out on the opportunity to realize the vision you started with. Speed without control means making so many compromises that you end up with a poor substitute for your idea.
The solution isn't to abandon AI-assisted development—far from it. The solution is to recognize that impressive demos are the beginning of the journey, not the end.
The final 20% can look different for every founder, but here's what it might include:
• Complex user flows: Beyond basic forms to multistep processes with conditional logic.
• Custom integrations: Connecting with the exact services your business relies on.
• Flexible data models: Adapting as your business requirements evolve.
• Performance optimization: Ensuring your app remains responsive as usage grows.
• Compliance requirements: Meeting frameworks like GDPR or SOC 2.
The solution is to choose a platform that will support your vision, and your business, long term.
My philosophy (and Bubble's) is that AI should amplify your vision, not limit it. The future belongs to platforms that see AI as a partner rather than a black box or wish-granting genie. Tools that expect humans to understand and engage with AI output, not just accept whatever it produces. This is why the combination of AI and visual development is the real future—they help you not just start fast but finish strong.
For founders, the ability to go from 80% to 100% defines success. Because the difference between having something to show and having something to launch, scale and iterate isn't just a technical distinction—it's the difference between an idea and a business.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?