Thought Leaders

The App Economy Is Changing: What Happens When Anyone Can Build Their Own Software?

Bindesh Vijayan
Add Unite.AI to your preferred sources on Google
Key PointsPersonal apps serve one person while production software carries accountability to…
View more
  • Personal apps serve one person while production software carries accountability to strangers who depend on it.
  • AI accelerates coding but 46% of developers distrust its accuracy according to the Stack Overflow survey.
  • AI amplifies existing team strengths so engineers shift from writing code to judging it.

From Choosing Apps to Describing Them

On October 13th 2026, Google announced that Android users can now build their own apps with Google AI Studio and run them privately on their own phones through Google Play, with no public store listing required. The apps are “vibe-coded”: you describe what you want in plain language, and the AI writes it. The first release is limited to the US and to users aged 18 and over, and sharing with a small circle of trusted contacts is planned for the months ahead.

The shift is simple to say. For fifteen years, the question was “Which app should I download?” Now it can be “What do I want my phone to do?” Sam Bright, Google’s VP and GM for Google Play and Developer Ecosystem, put it as a move from “there’s an app for that” to “there’s an app for me”.

Think of a parent who wants one screen showing this week’s school pickup rota, the football schedule and the grocery list. No store has that app, because no store would ever build it. Today, that parent can describe it and have it on their phone by the evening.

I welcome this. It puts real creative power in the hands of people who never had it. But it also invites a question worth answering clearly: is a personal app the same thing as software? In my view, it is not.

Building an App Is Not the Same as Engineering Software at Scale

A personal app might only need to work for one person. Software used by customers, employees or the public has to work reliably for people who depend on it. Both are software, but the responsibility changes when other people rely on what you build.

Personal-use application Production software serving multiple users
Who uses it You, or a few people you trust Hundreds to millions of strangers
Example A habit tracker, a family chore board, a recipe scaler A payroll system, a hospital booking platform, a bank’s mobile service
If it breaks You may be able to fix it yourself People are not paid, appointments are lost, money moves wrongly
Data May include personal or sensitive information Other people’s, held under legal duty
Security Privacy and security still matter Needs robust security, monitoring and appropriate audits
Connections Few or none Payments, identity, ERP, regulators, other systems
Life span May be short-lived or maintained over time Maintained, upgraded and supported for years

A useful test: ask what happens at 3 a.m. when it fails. For a personal app, you might be able to fix it in the morning. For a service that people rely on, someone may need to respond immediately.

Not every app needs the same level of engineering, and not every task should be entrusted to an untested app. A gym-timer you built for yourself is a personal app. A system that holds your members’ payments, health data and contracts is software, and it carries obligations that a prompt cannot replace.

Who Google’s Move Is Really For

Personal apps do not appear in Play search or listings and are not shared with anyone except their creator. Sharing with trusted contacts is still to come. A popular app can later be moved to a public listing, but only after trust-and-safety reviews, and apps must still follow Google Play’s policies. Google’s own announcement frames the update around Play.

That is a design for individuals, families and small circles. It is not a design for a company rolling out a system to its staff and customers.

Here is what that looks like in practice:

  • A student builds a revision planner that matches her exam timetable.
  • A shop owner builds a private stock-check screen for himself, so he can see what is low while standing in the aisle.
  • A cricket club captain builds a team availability poll for twelve friends.

These are meaningful, useful things, and until now most of them were never built because the cost was too high. That is the real gift here: the cost of building something small and personal has dropped to almost nothing.

What It Means for Existing Software Makers

Personalised apps will take some ground, and it is worth being clear about where.

Where they compete. Small, single-purpose utilities that people used to download because nothing else fitted: a custom unit converter, a simple expense log, a workout timer, a to-do list shaped exactly your way, a utility tool that you envisioned but didn’t know how to build. . Why pay a subscription for a tool that does 5% of what you need, when you can build one that does 100% of it?

Where they do not. Anything where trust, scale or accountability is the product: banking, payroll, healthcare records, logistics, identity. Nobody wants a one-off, unreviewed app holding their salary or medical history. Those systems need shared data, audit trails, integrations and people accountable when something goes wrong.

So the honest picture is not “apps replace software.” It is that the long tail of tiny needs finally gets served, while serious software is judged on what it was always meant to deliver: reliability, security and support. Incumbents who rely on being merely convenient should pay attention. Those who rely on being dependable have a clearer story than ever.

Reliability Is Built In, Not Bolted On

AI makes writing code fast. It does not make code correct. The 2025 Stack Overflow Developer Survey found that 84% of developers use or plan to use AI tools, yet 46% say they distrust the accuracy of AI output, against 33% who trust it. Even the people using these tools every day check the results.

For a simple personal app, the consequences may be manageable, although privacy and security still matter. If your habit tracker miscounts a day, you fix it. For software that serves other people, the same slip can mean a leaked record or a wrong payment.

That is why serious teams treat quality as part of how software is made, not a final polish. The NIST Secure Software Development Framework describes this: practices for preparing the organisation, protecting the code, producing well-secured software and responding to vulnerabilities.

The challenge grows as an app changes. A new feature can affect existing functions, data structures or access permissions. Without testing and version control, a small change can break something that previously worked.

An everyday comparison: a home cook can improvise a dinner for the family. A restaurant needs hygiene checks, supplier records and a process that works the same way every night. Both are cooking, but only one is held to a standard.

Owning What You Build

When you build something, “ownership” means three different things. It helps to ask all three questions before you commit.

  • Code ownership: can I take the source with me? Example: you build a budgeting app and later want a developer to extend it. Can you export the code, or does it live only inside one platform?
  • Data portability: can I take my information with me? Example: after a year of logging expenses, can you export them as a file you can open elsewhere, or are they locked in?
  • Application portability: can the app run outside the platform where it was built? Example: if you need to change hosting providers or move to another development environment, can you take the app and its dependencies with you, or must it be rebuilt?

A personal app that is easy to build but hard to leave can become a quiet trap. Google’s private distribution model limits who can discover these apps, but privacy also depends on how the app handles data and permissions. The wider lesson holds for any AI builder: understand what you can export, what depends on the platform and what happens if you want to leave. Choose tools that let you keep your code, your data and your options.

Why Engineers Matter More, Not Less

It is tempting to think that if AI writes the code, engineers become less important. The evidence points the other way. Google’s DORA 2025 research describes AI’s primary role as an amplifier: it magnifies an organisation’s existing strengths and weaknesses.

A team with good testing, clear ownership and sound design gets faster and better. A team without them gets faster at making mistakes. The biggest returns come from the surrounding system, not the tool itself.

Picture two teams given the same AI assistant. One reviews every change, tests it and monitors it in production. The other ships whatever the AI produces. Within a few months, the first has a product customers trust, and the second has a pile of problems nobody understands.

So the engineer’s job is moving from typing code to judging it: deciding what to build, checking what the AI produced, and keeping the whole system safe. That is a more valuable skill, not a lesser one.

The Next Chapter

Google’s announcement is good news. Anyone with an idea can now turn it into a working app on their own phone without any prior experience or knowledge of coding, and only by specifying the intent of what they want to build, without hiring a crew of human specialists. That widens who gets to build, and it will produce many small, useful things that never existed before.

The lines worth keeping clear are these. A personal app can serve one person’s needs. Software that other people rely on carries greater expectations of reliability, security and support. Both deserve to exist, and each needs a different level of care.

So here is my invitation. If you have ever thought “I wish there was an app for this,” stop wishing. Why not build one?

Bindesh Vijayan is the Co-Founder and Chief Technology Officer of Myndlab, where he leads the design and development of advanced AI systems that turn ideas into deployable, production-ready software. With more than 14 years at Microsoft, including participation in the beta testing of what would become Microsoft Azure, Bindesh brings a plethora of experience across some of the most important technology shifts of the last three decades, from early computing to cloud and now AI-native development.

His broader career includes roles across major global organizations. From Fortune 500 enterprises including SAP, and Infosys, to global strategy powerhouse Accenture, where he worked at the intersection of engineering, enterprise systems, and business transformation. Known for combining hands-on technical depth with strategic vision, he has built and led high-performing engineering teams across multiple markets and technology environments.

At Myndlab, Bindesh is focused on shaping the next era of autonomous software delivery. His expertise spans large language models, AI agents, prompt engineering, model optimization, and production-grade AI architecture, with a focus on building systems that are scalable, efficient, and commercially useful. He is particularly interested in how intelligent agents and end-to-end AI workflows can close the gap between concept and execution, making advanced software creation faster and more accessible for businesses and developers.