When the iPhone SDK and Apple App Store launched in the late 2000s, I was just starting my career as a software engineer. I had grown up programming for fun and watching the internet mature through the 1990s, and this felt like the moment everything snapped into place. This was the next big thing. Everyone could feel it. Everyone was paying attention.
Being an iOS developer felt like being a rock star. Everyone had an app idea. People shared apps the way they used to trade baseball cards. Apps felt magical and fun. Apple didn’t just give us a new way to download software. They gave us a new way to interact with the world, and they made it frictionless.
In the early days, apps were primarily novelty items: a virtual mug of beer – one you could ‘drink’ – a lightsaber or digital Zippo lighter. Over time, that playful experimentation gave way to something much bigger. Banks showed up, along with payments, navigation and identity. Social platforms like Facebook, Twitter, and Instagram took off once they went mobile. The smartphone became the primary interface to the digital world, and the App Store was how you accessed it.
As apps matured, they stopped feeling exciting. They became infrastructure. Once that happened, it became incredibly hard to displace. But history shows it can happen, and it tends to happen the same way: a new interaction model arrives, and suddenly using a legacy interface feels like extra work.
How the App Store model shaped digital behavior
Anyone who has lived through the era of the early App Store likely feels nostalgic about it, and many of us have been waiting for another similar moment in time. AR and VR made a splash. Crypto and NFTs had their time in the spotlight. But none of these truly reshaped our daily behaviors.
AI feels different. You can’t avoid hearing about it. You can’t really avoid using it anymore either. And like the App Store before it, AI is starting to change how we behave. How we ask for things. What we reach for first when we want to get something done.
Apps trained us to think in interfaces. You opened the right app and navigated a polished, animated environment to get what you wanted. AI flips that around. Instead of thinking about interfaces, we think about outcomes. Book this. Plan that. Fix this. Find me the best chair that matches this picture of my living room, within this budget, with these constraints.
For those of us who lived through the mobile app era, today’s pattern with AI feels familiar: exciting, uncertain and messy.
Instead of searching for things, we ask for what we want and let the system figure it out. Instead of editing text or images ourselves, we describe what we want changed. Planning trips, ordering groceries, managing playlists now happens inside AI assistants like ChatGPT without opening the individual apps we once relied on, because those services increasingly live as apps inside ChatGPT itself.
That’s why OpenAI’s push to build an app ecosystem inside ChatGPT matters. It’s not that these integrations are perfect today. They’re not. You can’t always do everything you could do in the native app. Sometimes it takes longer. Sometimes you hit errors. Getting the right prompt can be confusing. In many ways, this feels a lot like the early App Store days: powerful ideas, experimentation and rough edges.
But the direction is clear. A growing share of consumers say they’ve already replaced, or expect to replace, at least one traditional app with an AI assistant. Just like with apps, as users begin to adopt a new primary interface, brands must meet consumers where they are, even if it isn’t fully baked yet.
What it means for users, platforms, brands, builders
For users, app fatigue is real. Sitting down at a restaurant and being told to download another app, create another account, and hand over more data just to order food is frustrating. A single conversational front door feels simpler. If I’m already in ChatGPT for half a dozen other tasks, it’s natural to wonder why I can’t do more.
At the same time, new questions emerge. What can this actually do? How do I ask it the right way? Can I trust it with sensitive data? Can I rely on it to get things right? In the app era, the interface guided you. In the AI era, users have to learn how to ask.
For platforms, the stakes are especially high. Whoever controls the entry point holds enormous leverage. Apple is still very much a major player here. These AI interactions are largely happening on iPhones. Apple’s hardware, privacy posture, and on-device AI capabilities position it well, even if the App Store itself evolves. It’s easy to imagine deeper integrations where AI assistants and operating systems work together, rather than one cleanly replacing the other.
AI platforms offer massive reach, much like the App Store did. But they also capture margin and control discovery. There’s real platform risk and real opportunity at the same time. The danger isn’t adopting AI too early. It’s pretending this shift isn’t happening at all.
When it comes to brands, the conflict isn’t about apps vs AI. It comes down to discovery. In the mobile era, discovery meant App Store rankings and search. In the AI era, it means something different: When a user asks an AI assistant what to do, what to buy, or what provider to trust, is your service, product or brand mentioned? Does it direct the user to your service? Or does it complete the task without your brand ever being surfaced?
For builders and software providers, the parallels to the early mobile age are hard to ignore. Back then, teams scrambled to expose data through APIs that worked well on phones. Now the scramble is to make data usable and reliable for large language models. This is as much a design challenge as a technical one. How do users discover what your service can do through conversation? How do you make sure the AI actually completes the task the user intended, safely and accurately?
Key takeaways for builders are starting to look clear:
- Design for conversation, not navigation
- Make your service usable through APIs and agent workflows, not just through user interfaces
- Treat trust, accuracy, and guardrails as product features, not afterthoughts
For those of us who lived through the mobile app era, today’s pattern with AI feels familiar: exciting, uncertain and messy. The technology isn’t perfect, but that’s expected.
The biggest interface shifts never arrive polished. They arrive early, disrupt the status quo, and reshape everything as they mature. However, since we know how this story ends, we should embrace it. History shows us that transformative technologies aren’t judged by their early imperfection, but by the possibilities they unlock. AI is no different.
The only question that remains is: what will you build?
Be First to Comment