Before You Commission a Mobile App: What AI-Assisted Development Changes (and What It Doesn't)

If you've been getting quotes for a mobile app recently, you've probably noticed the spread is enormous — a few months and €20k on one end, a year and €150k on the other. Some of that gap is scope. A lot of it is disagreement about what AI-assisted development actually changes. This post is about that disagreement, from someone who builds mobile apps for a living with these tools.

The Cross-Platform Question Is Now Settled

For most of the past decade, there was a real argument for building native iOS and Android separately. The tooling had gaps, the performance was noticeably worse, and the ecosystem was immature. That argument is gone. React Native with Expo is what Shopify, Discord, and Coinbase ship — not because they couldn't afford native, but because one codebase targeting both stores with native-quality results is simply the better engineering decision for the vast majority of business apps.

What AI-assisted development removes is the last remaining counterargument: that cross-platform has a velocity penalty. It doesn't anymore. When planning, boilerplate, and API integration are all AI-accelerated, the setup work that used to eat two to three weeks shrinks to a few days. The codebase you start with is cleaner than what a traditional dev shop would hand-write in the same time.

What AI Actually Compresses

The speed gains are real, but they cluster in specific places. Navigation structure and routing setup — the scaffolding that every mobile app needs and that takes the same amount of effort regardless of what you're building — used to be slow, careful work. With AI-assisted tooling it's a starting point generated and reviewed in hours, not days. The same goes for typed API contracts between the app and the backend, auth flows, offline data caching patterns, and CI/CD configuration for both stores.

On a recent project, we had a working build running on both iOS and Android simulators, connected to a real backend, with authentication and the core navigation in place inside five days of development. That's not vibe coding — every piece was reviewed, tested, and gated by CI. It's just that AI compresses the parts of mobile development that are tedious and well-understood, which turns out to be a significant fraction of early-stage work.

Where Human Expertise Still Decides the Outcome

App Store and Play Store submission is not a solved problem that AI makes easy. Apple's review guidelines are specific, frequently updated, and full of underdocumented edge cases. A reviewer will run your app on a real device, tap in unexpected sequences, and reject it for something your simulator never showed. Handling that process — understanding the rejection reasons, knowing when to push back and when to adjust, getting through review on the first or second attempt — is accumulated experience. No amount of AI tooling substitutes for it.

The same applies to native module integration: biometrics, background location, push notifications, camera pipelines. These require working with platform-specific APIs that change with OS updates, have different behavior on different device generations, and fail in ways that only show up on physical hardware. A developer who has shipped several apps has seen those failure modes. One who hasn't doesn't know what they don't know.

The Vibe Coding Trap Hits Harder on Mobile

When a vibe-coded web app breaks, it usually breaks in production, quietly — a log somewhere, a support ticket, a user who doesn't come back. When a vibe-coded mobile app breaks, it breaks publicly. A 1-star review with "crashes on startup" is permanent. An App Store rejection means days of delay, a follow-up review request, and often a rewrite of the component that failed. Apple reviewers are not sympathetic to demo-grade code.

AI-assisted development produces fast output. Vibe coding produces fast output that skips the review, the tests, and the CI gate. The difference between the two is entirely about what happens between the AI generating code and that code reaching the store. At Rebooted Solutions the answer is the same as it is for web projects: every PR is reviewed, every feature has tests, CI gates are non-negotiable. The speed comes from AI compressing the right things — not from skipping quality gates.

Three Questions to Ask When Evaluating a Mobile Build Partner

• Have they actually handled App Store submission? Ask them to name a specific rejection reason they've resolved in the last twelve months — what the issue was, how Apple described it, and what they changed. Anyone who has gone through real submissions can answer this concretely. Anyone who hasn't will give you a vague answer about "following guidelines."
• Can they show you their CI setup and test strategy? A partner who ships quality mobile code has an automated pipeline — unit tests, integration tests, device matrix builds — running on every pull request. If the answer is "we test manually before release," that's the signal that AI speed and human review aren't both actually present.
• Who owns the codebase after handover? The app should live in your repository, in your App Store developer account, deployed from your CI pipeline. If the agency controls the certificates, the store accounts, or the deployment keys, they've created a dependency. That's a negotiation you're going to have when the relationship ends.

Mobile development in 2026 is genuinely faster than it was three years ago, and the cross-platform tooling is mature enough that most business apps don't need to pay for two separate native builds. The gap between a partner who uses AI as a discipline multiplier and one who uses it as a shortcut is invisible in the quote but visible in the App Store within months of launch.

Rebooted Solutions builds cross-platform iOS and Android apps with React Native and Expo — AI-assisted at every step, with code review, CI/CD, and full App Store submission handled. If you're evaluating whether a mobile project is the right next step or comparing approaches, book a discovery call and we'll give you a straight technical assessment.