Skip to content

// essays · 30 September 2025 · 3 min read

Why every app will become an AI app

I’ve started noticing a pattern.

Every time I use a new product, I half expect there to be an AI button somewhere. If it’s not there, the app feels unfinished, like it forgot the most obvious feature.

This isn’t because I’m obsessed with AI. It’s because once you’ve seen how much drudgery AI can remove from a workflow, you can’t help but notice the absence.

Apps used to be about giving you a set of tools. You opened them, you did the work, and you got the result. Now they’re shifting into something stranger. They don’t just give you the tools, they try to do the work alongside you. Sometimes they even try to do the work before you ask.

That’s a fundamental change.

When people talk about AI apps, they usually point to flashy demos, writing essays, generating images, building websites. But the real shift is quieter. It’s happening in workflows.

A workflow is just a set of steps you repeat to get something done. You open a slide editor, you gather notes, you draft, you polish. Most of those steps are boring. Most are mechanical. That’s where AI slides in.

At first it’s small: a summarizer, a draft generator, a suggestion. Then it starts planning. It starts anticipating. Eventually the app doesn’t feel like a tool at all. It feels like a co-worker hidden behind the interface.

The cost of building this in is dropping to almost nothing. A few years ago, adding AI meant a research team and racks of GPUs. Now it’s an API call. You add a retrieval layer for context and you’re done.

When the cost of something goes to zero, it spreads. Not just to one category of app, but to all of them. The question isn’t which apps will become AI apps. It’s whether any app can survive if it doesn’t.

What makes this even more inevitable is users. Expectations rise quickly and never fall back.

Search engines used to be keyword boxes. Now they feel broken if you can’t just type a natural question. Project managers used to expect you to write your own updates. Once you’ve used one that drafts them for you, everything else feels like homework.

We don’t go back to worse workflows once we’ve seen better ones.

Of course, AI isn’t perfect. It makes mistakes. It hallucinates. It can break trust if you let it run wild. But these problems aren’t reasons to ignore it. They’re design challenges. You add guardrails. You build in retrieval so it’s grounded. You let humans override it when it matters.

The bigger risk is ignoring it. If your app doesn’t learn how to take away the dull parts of a workflow, someone else’s will.

The funny thing is that in a few years, we won’t even talk about “AI apps.” They’ll just be apps. And the ones that didn’t adapt will look like relics from a slower, clunkier time.

The future isn’t that every app gets an AI button. It’s that the button disappears, because the whole workflow has been reshaped around intelligence.

That’s why I think every app will become an AI app. Not because it sounds good on a slide deck, but because users will demand it. Once you’ve seen an app do part of the work for you, you’ll never want to go back to doing all of it yourself.

Muhammad Mairaj is a Senior Backend Engineer with 8 years building payment and data systems. He writes about data engineering, measured. More about him.