Mobile App Trends – Insights from Perth's Top Developers

Most of what a Perth business asks for in a mobile app has not changed since 2024. What has changed is how cheaply it can be delivered, and how much of it you should not build at all.

This post is our working list of the trends that survived contact with real projects, plus the ones we were wrong about, with a note on what each costs.

Build once, ship to both stores

Cross-platform is no longer a compromise. We build with Vue and Capacitor: one codebase, wrapped into an iOS and an Android app, with native access through plugins where it is needed.

That approach covers the large majority of business applications. It stops being the right answer when the app leans hard on native performance, background processing or deep hardware access, where a fully native build still wins.

The practical effect is on budget. A single codebase means one set of tests, one release process and one team, which is usually the difference between shipping this year and not.

Progressive web apps took the low end

If the app does not need to sit in an app store, a progressive web app is usually the better first release. It installs to the home screen, works offline with a service worker, and updates without a review queue.

What it does not do as well: background location, push notifications on iOS, and anything the platform reserves for native. Those limits are narrower than they were in 2024 and still real.

For a customer booking something, checking a status or reading a record, a PWA is usually enough. For a field app that must stay awake all day, it is not.

AI moved from feature to utility

In 2024 "AI integration" meant a chatbot. What actually earns its place now is narrower and more useful:

• Classifying an incoming enquiry so it reaches the right person.
• Summarising a long document into the four facts someone needs.
• Extracting structured data from a photo of a form or an invoice.
• Drafting a first reply that a person edits.

Each of these replaces a specific task. Each is cheap to try and easy to measure. The failure mode is a general assistant bolted onto a product that nobody asked to talk to.

5G, AR and blockchain: the honest scorecard

We listed all three as emerging in the original version of this post. Two years on:

• 5G. Real, widely available in Perth, and rarely the reason an app works or does not. The gains show up in live video and large uploads; everyday apps felt no difference.
• AR. Genuinely useful for a handful of jobs, such as equipment guidance and site walkthroughs. Almost no Perth small businesses have one of those jobs.
• Blockchain. Still waiting for a business case outside of specialist finance work. We have not built one.

IoT is the one that quietly arrived. Sensors on equipment, tanks and cold rooms now feed software that did not exist as a product when this post was first written.

Security is the trend that never stops

Every year the same list: encrypted transport, hashed credentials, multi-factor sign-in, no secrets in the app bundle, an update path for dependencies.

The change since 2024 is regulatory rather than technical. Australian businesses now have clearer obligations about the personal information an app holds, and a breach is expensive in a way it was not a decade ago. An app that stores customer data is a target whether or not the data looks valuable.

If you take one line from this post, take that one. The rest are options.

What this means if you are commissioning an app

• Start with a web app or PWA. Move to native when a specific feature forces it.
• One codebase, unless the product genuinely needs two.
• Add AI to a task, not to the product.
• Budget for maintenance from year one. Roughly 15 to 20 per cent of the build cost each year is a reasonable rule of thumb.

If you want a second opinion on a scope before you commit to a build, send it over. We will tell you what we would cut.