How AllTrails Scaled AI Coding Adoption to 90+% and Accelerated Testing with DevClarity

AllTrails is a globally-known outdoor discovery and navigation platform that connects a worldwide community of trailgoers to vetted routes and on‑trail guidance. Through its app and website, it helps people find new places, plan adventures, and elevate their experience outdoors—from weekend hikes to trail runs and rides. Founded in 2010 and headquartered in San Francisco, the company sits at the intersection of technology and outdoor recreation.

By mid-2025, AllTrails had a promising start with AI adoption—developers were experimenting with tools like GitHub Copilot & Cursor. But as engineering leaders looked ahead, they saw a critical gap: the team lacked the repeatable workflows and systematic approach needed to maximize and sustain AI coding impact.

Additionally, as AllTrails began to accelerate feature velocity with AI coding tools, it needed to maintain test frequency and ensure continued quality. Ultimately, the team wanted sustainable, AI-native workflows that the entire engineering organization could adopt.

AllTrails partnered with DevClarity to train the engineering team on how to leverage AI coding tools most effectively. DevClarity also came alongside AllTrails as an embedded team, working directly inside their Android and iOS codebases, to accelerate specific high-impact projects. The engagement combined foundational work—creating context rules and reusable prompts for AllTrails' repositories—with hands-on training and focused directed efforts on mobile testing.

The approach was deliberately pragmatic. Rather than abstract training sessions, DevClarity conducted "laptop‑open" working sessions where AllTrails developers wrote real tests for real code. The team established Cursor as their primary AI code editor, with Copilot and Claude filling platform-specific gaps, and built custom context files that captured each platform's testing conventions.

The process included:

• • Foundational AI Coding Training: The entire engineering team engaged in training sessions tailored for AllTrails to set the foundation for AI-native coding.
• • Android workflow optimization: Built AI-native testing with JUnit 5, Truth, and MockK; discovered a feature‑module strategy that dramatically reduced build dependencies and iteration time
• • iOS iterative refinement: Working sessions with the iOS team to solve data mocking challenges and establish human‑in‑the‑loop checkpoints that prevent AI from skipping critical steps
• • Unified four-step pattern: Both platforms converged on a create → analyze and plan (with human verification) → write → run workflow, balancing AI automation with developer judgment

The numbers tell a compelling story. AI tool adoption jumped to +90%, with the "not using AI" segment dropping to just 10%. Developers reported that AI accelerated 40% of repetitive tasks and lifted overall output by 20%.

But the real impact shows in the workflows themselves.

• • Both Android and iOS teams now have repeatable, documented patterns for AI-assisted test writing that they can run independently.
• • The Android team's feature module strategy delivers faster build cycles with fewer dependencies.
• • The iOS team's refined prompts generate clean with great structure and appropriate mocking.

With company-wide training, systematic context creation, and real-world AI application within their codebases, AllTrails has set the stage for an AI-native engineering team and many more amazing features to come!