Every week, I watch apps get built that nobody would have budgeted for
AI doesn't create audiences or needs. It makes the apps that serve them worth building, and they're multiplying before our eyes.
An audience can be too small to justify custom development, and still big enough to sustain an app.
For a long time, the first half of that sentence decided everything. Before anyone asked what service an app could deliver, you needed a budget, a team, and specialized skills. Many projects ended right there. Not because their audience didn’t exist, but because the cost of the channel was out of proportion to its size.
App builders changed that equation a first time. They brought design, content management, mobile features, and publishing together in a single tool. An organization could ship an app without becoming a software company itself.
AI is now lowering that threshold on new fronts. It makes more specific features possible, pulls more utility out of existing content, and simplifies some of the daily work of running an app. It doesn’t create audiences or needs. It makes the math work for more of the apps designed for them.
That, as I see it, is where abundance will come from: not a single market flooded with lookalike apps, but a multiplication of projects built for specific audiences.
Narrow from a distance, viable up close
Every week, I answer support tickets from our customers. That’s where I’ve seen a news outlet covering lower-division Argentine soccer, and a fashion magazine built for a Roblox community. From a distance, those subjects look narrow. Up close, they generate enough content, enough reasons to come back, and enough habits to sustain an app.
These projects didn’t start with AI. They’ve long been the everyday reality at GoodBarber and at the other tools that made building apps more accessible. They’re a reminder that a market doesn’t need to be huge to be coherent. It needs an identifiable audience and a service useful enough for that audience.
An app about Argentine soccer isn’t trying to beat a major sports network on its own turf. A magazine made for a Roblox community isn’t competing with the fashion press. A members-only space for a nonprofit isn’t chasing the general public.
So these apps aren’t all competing for the same attention. Some want to build an open audience. Others serve customers, readers, or members they can already name. Multiplying them doesn’t necessarily multiply competitors in one market. It gives more specific markets a tool built to their size.
Abundance can produce noise. It can also produce sharpness.
Abundance starts with the attempt, not the launch
On Google Trends, relative interest in the search “create an app” has roughly doubled between its 2023 plateau and the levels seen in 2026. That curve doesn’t count future creators or future apps. More modestly, it shows that building an app is taking up a growing place in what people search for.
Over the same period, Google Play’s catalog fell from 3.4 million apps in early 2024 to about 1.8 million by April 2025, after Google tightened its quality rules.
Neither number proves that more useful apps will get published. What they do is locate the movement. Searching for how to create an app, producing a first version, and publishing a product users will keep are three different steps. Interest and attempts can climb while publication stays selective.
I’ve made the case before that the rise of app generators is good news for app builders. Their contribution isn’t limited to the code they produce. They move more people from “that’s not for me” to “I could try this.”
That’s where abundance begins: in the number of needs that get a first working form, even if not all of them will become published products. For niche projects, that attempt matters more than most. Their economics rarely allowed a long design phase before finding out whether the service would find its audience.
A niche app comes down to three things, and AI is moving all three
None of this means every store, creator, or nonprofit should have an app. A newsletter, a website, a video channel, or a private group can still be the better channel.
Democratization doesn’t make an app automatically relevant. It means the channel gets judged on the service it can deliver, instead of being ruled out before anyone has looked at that service.
For a project aimed at a specific audience, three things decide whether an app is justified: what you can build into it, how useful it is to its audience, and how much effort it takes to run. AI is now acting on all three.
When a project needs a specific feature, its creator can describe a custom section in plain language, test it inside the app, and refine it through conversation, without calling a developer first. Customization becomes an option for needs that would never justify bespoke development.
A publisher can make its archives searchable in natural language through a RAG chatbot that answers only from its own content. The app doesn’t hold more articles. It gives a specialized audience a new way to tap what’s already there.
A merchant can use an assistant connected to the app through MCP to create a product, add it to a collection, and schedule a notification, without clicking through each step in a dashboard. For a small operation, shaving that daily effort matters as much as cutting the cost of version one.
These three don’t do the same job, but they improve the same equation. They cut the effort it takes to customize an app, make it more useful, or run it day to day.
Substance before interface
I watched a couple gather and organize more than a thousand texts, then add a conversational interface on top. A reader can now ask a question in their own words and surface the relevant passages from the collection. That way of exploring it would have been out of reach for its creators until very recently.
AI made the collection easier to access. It didn’t produce the texts, the selection, or the editorial judgment that holds it together.
A prompt can produce an interface. It can’t produce what makes it useful. In a content app, that’s a body of work and an editorial line. In a community app, it’s a relationship someone maintains. In a commerce app, it’s an offer, earned trust, and customers with a reason to buy again. In a service app, it’s a real-world task that mobile makes simpler or more direct.
There’s still the work of turning a first version into an app that’s published, reliable, and built to last: meeting store requirements, protecting users, maintaining the product, and keeping the promise that earned the download.
The idea and its first draft are getting cheaper. Relevance, execution, publication, and staying power are still what earn a niche app its place.
Choosing an app for the right reasons
At GoodBarber, we’ve been working since 2011 to put mobile technology within reach of people who have neither a development team nor a big-brand budget. Our paying customers are in 152 countries today. What interests me about that number isn’t scale. It’s the vantage point: when a tool becomes accessible, it gets used for projects, cultures, and audiences its maker could never have predicted alone.
AI is now widening the pool of projects that can clear that threshold. Their creators show up with content, an offer, a community, an expertise, or a service. An app builder can carry the shared layers of design, publishing, and operations. AI can make the parts specific to their audience more attainable.
The outcome won’t be that everyone gets an app. It will be that far more people get to choose one for the right reasons, instead of giving up on it for the wrong ones.
That’s what the coming abundance looks like: not apps all chasing the same crowd, but each one worth building because it’s useful enough for the people it serves.