# Shop App Successfully Migrates from React Native to Native Swift/Kotlin AI-assisted migration completed in 12 weeks with significant performance gains - **Source:** [Shopify Engineering](https://shopify.engineering) - **Published:** 2026-09-10 - **Topic:** [Software & Engineering](https://brifing.si/c/engineering) - **Original article:** - **On Brifing:** https://brifing.si/s/2166876f-1bbc-4204-b5d6-6b9d00e094c8 ## The Migration Decision Shopify's engineering team faced a pivotal choice: continue with React Native or transition to native development. With the Shop app serving hundreds of millions of users since 2020, the team recognized that advances in AI coding agents made native development more feasible. The decision coincided with Shopify's planned React Native New Architecture investment, creating an opportune moment to evaluate alternatives. A proof-of-concept demonstrated that feature-for-feature migration was achievable within tight timelines. ## Proof of Concept Success Before committing fully, one engineer spent a week testing AI-assisted migration. Using React Native as reference, they successfully recreated screens, interactions, and flows in SwiftUI. While not production-ready, this experiment proved that complete migration was possible. The POC showed agents could effectively port features, scaffold screens, wire data, and implement animations when working from existing implementations. This gave the team confidence to proceed with full migration. ## Migration Team Structure A core group of six engineers built the native foundations and main user journeys. Feature teams joined later to validate specific areas and handle edge cases. The team prioritized maintaining both feature behavior and analytics events that downstream systems depended on. They ensured continuity for users - preserving sign-in status and push notifications while deliberately simplifying some app elements. The migration was treated as a normal update to minimize user disruption. ## Performance Improvements Native versions showed dramatic improvements across key metrics. Session stability increased from 99.5% to 99.95% - a 10x reduction in crashes. The Android build size decreased by 109MB while iOS remained similar. Android build times fell approximately 75%. The native Android app achieved 120 FPS during scrolling and navigation on Pixel devices, demonstrating significant rendering improvements with minimal optimization efforts so far. ## AI Development Workflow The team implemented novel agentic workflows differing from traditional development. Multiple agent sessions ran across separate worktrees simultaneously. Engineers focused on providing clear tasks, maintaining fast build-test loops, and frequent change reviews. They built a reusable migration workflow as an extension for their Pi coding agent, with specialized subagents handling different migration aspects from source inspection to parity reviews. ## Tardis Debugging Tool To support agent development, the team created Tardis - a debugging tool giving structured access to live app events, logs, and state. Agents could investigate issues, check navigation/analytics, and validate fixes without manual UI inspection. Tardis included comparison features capturing screenshots and events from both React Native and native apps at checkpoints. This enabled comprehensive parity verification while accounting for expected differences like timestamps. ## Implementation Governance The team implemented robust governance for AI-assisted development. Engineers reviewed requirements and plans before implementation. Plan acceptance was cryptographically tied to content hashes - any changes invalidated previous approvals. This ensured reviews matched the actual implementation. The process covered UI, state, navigation, analytics, accessibility, and data behavior, maintaining quality while enabling rapid development. ## Migration Impact The successful 12-week migration demonstrates how AI can transform complex technical projects. Beyond immediate performance gains, it establishes a template for future platform transitions. The project highlights how coding agents can change cost-benefit calculations for maintaining shared codebases versus platform-specific development. Shopify's experience provides valuable insights for other large-scale mobile applications considering similar transitions. ## Sources covering this story - https://shopify.engineering/shop-app-migration — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._