JamBoard: how AI helped us build our app for musicians (without doing it for us)

Neto Pompeu —

JamBoard: how AI helped us build our app for musicians (without doing it for us)

On May 14, 2026, we shipped version 1.1 of JamBoard on TestFlight. An iOS app built for music bands — to manage rehearsals, vote on dates, track gig fees. Six months of development, two Apple rejections, a complete rewrite of the authentication flow, and intensive use of AI throughout.

This article is an honest retrospective on what AI really enabled us to do — and, more importantly, what it doesn't do and that we had to master ourselves for the project to ship.

The project: an app for music bands

JamBoard was born from a simple observation: bands in 2026 still organize themselves with overloaded WhatsApp threads, Google Forms polls, shared Excel files. Nobody had built the app that brings it all together — rehearsal scheduling, date voting, gig fee tracking, score sharing, internal communication.

We built it. The scheduling agenda is 100% free, with a Premium tier (€1.99/month or €12.99/year) for rehearsal date voting and gig fee tracking. Target audience: amateur and semi-pro bands. Available in 5 languages from V1 — French, English, Mexican Spanish, Brazilian Portuguese, German.

What AI really sped up

Let's be precise. On the ~3,500 hours of work estimated for this type of app, we probably saved 30 to 40% thanks to AI. Here's where the gain was real.

Swift and SwiftUI boilerplate

All view structures, view models, data flows between Firebase and SwiftData, error handlers: 80% of this code is repetitive. Claude Code and Codex CLI generate those foundations in minutes. The time saved is spent on the business logic that actually matters — where the value lies.

Multilingual translations

5 languages from V1 = 5× the localization work. Without AI, that's weeks of work with human translators. With Claude API + human review, it's a few hours. Quality stays excellent because we re-read every string in context, and we adapt nuances (Brazilian vs European Portuguese for example).

Unit and UI tests

Writing tests by hand kills morale — and ends up being the first thing teams skip. AI generates 90% of tests from production code. We verify, we adapt edge cases, and coverage rises effortlessly. By V1.1 ship, we had an automated test suite running on every commit.

Debugging and Apple rejections

When our build 45 was rejected for a Sign in with Apple bug on iPad and a misconfigured Firebase demo account, Claude analyzed Apple's report, reproduced the bug, and suggested three fix paths. We decided, we fixed, we resubmitted in 48 hours. Without AI, it would have been a week — time to dig through Apple's docs, retest on every device, write the submission notes.

What AI doesn't do — and that you must know how to do

This is where it plays out. AI is an excellent executor, but it does not decide for you. And in app development, technical and strategic decisions are 80% of the job.

Understanding App Store Guidelines

Apple has 14 categories of rules, hundreds of sub-rules, evolving interpretations. AI can read them, but it doesn't know what Apple will actually block this week. Our rejection on Guideline 2.1(a) — non-functional demo account — no AI had flagged it upstream. You had to know the review process pitfalls to avoid it.

Designing the right UX/UI

A semi-pro musician organizing a rehearsal at 10 PM from their phone — that's an experience to design. AI can generate ten mockups in five minutes, but a human has to know which one works in real life. App design is user empathy — not algorithmic optimization.

Choosing the technical architecture

SwiftData local-first or Firestore cloud-first? Sign in with Apple mandatory or optional? RevenueCat or native StoreKit 2? All these structural decisions commit the next five years of maintenance. AI suggests, but experience decides — otherwise you end up with technical debt that costs 10× more to fix afterwards.

GDPR, LGPD, CCPA

An app available in 175 countries means 175 regulations to know. The app must declare its data processing, present a DPA, respect the rights to erasure, portability, access. AI can summarize the texts, but it doesn't sign your Apple Privacy Manifest for you — and you are the one legally responsible before the CNIL or equivalent authorities.

Why a solid foundation remains essential

AI, in 2026, is a multiplier. But it multiplies your current level. If you have a strong command of iOS development, App Store Guidelines, product design and GDPR, AI will double your speed. If you start without those foundations, AI will produce code that compiles but gets rejected by Apple, crashes in production, and breaks the law.

This is precisely why training upfront remains essential. AI is not a shortcut to avoid learning — it's an accelerator for those who have learned.

Our approach at Primo-Studio

At Primo-Studio, we build iOS apps for our clients (and our own apps like JamBoard, PixConsent, AutoCare) by combining:

  • Over 8 years of digital experience — websites, mobile apps, digital marketing, audiovisual production
  • iOS technical expertise: SwiftUI, SwiftData, Firebase, RevenueCat, CloudKit, Sign in with Apple, App Store Connect API
  • Mastery of the App Store process — and the pitfalls that cost weeks (demo account, Privacy Manifest, SIWA, IAP, App Privacy)
  • Intensive but disciplined use of AI in our daily workflows, without ever delegating key decisions to it

If you have an iOS or Android app project for your business in French Guiana, we can support you — from full development to audit/consulting on an ongoing project, or training for your tech teams. To go further on AI in business, see our AI training guide for businesses in French Guiana.

Who is behind Primo-Studio

Néto Pompeu, founder. I grew up in Saint-Laurent-du-Maroni. Apprenticeship studies in mainland France, MBA in Digital Marketing at EDC Paris. I did my apprenticeship at Siemens, where I helped create the digital marketing hub of the Siemens Digital Industries subsidiary and worked on actions across France.

Returned to French Guiana in 2022 post-Covid, founded Primo-Studio SASU in 2024 with a clear goal: bringing European quality to the Guianese territory.

Primo-Studio is the first marketing agency headquartered in Saint-Laurent-du-Maroni. We serve all of western French Guiana — and far beyond: we've delivered projects across French Guiana, in mainland France, and for international clients.

Got an app project?

Contact us to discuss your project. Whether you're at the idea stage, mid-spec, or stuck mid-development, we'll look together at where we can accelerate — and where AI can (or cannot) help you.