How we Scaled AI Across our Mobile Engineering Team

We rolled out AI coding tools for writing code. Seven months later, engineers use them just as much for design docs, test plans, and PRD reviews.

The biggest gains weren’t where we expected.

Yes, engineers write code faster and write more tests. But the real shift was upstream. Increasingly clear requirements. Fuller test plans. Design docs that get pressure-tested before the first line of code. Time saved early compounds into cleaner execution downstream.

Here’s what we’ve learned.

The Starting Point: Support Without Mandates #

We build a mission-critical mobile platform used by field technicians daily. When AI coding tools gained traction, we wanted engineers to explore them while keeping quality steady.

The message was clear. Explore new tools. Push them. Leadership will support that work. We also addressed job anxiety early. AI was not replacing engineers. Engineers remain at the center of the work and validate the output. That clarity encouraged experimentation.

Why We Needed Better Tools for Mobile Workflows #

Mobile engineering has unique constraints. Xcode and Android Studio create specific workflow requirements. We needed industry-standard tools that fit native mobile development patterns and could be adopted across teams without custom integration work.

When we evaluated third-party AI coding assistants built for these environments, adoption was immediate. Engineers saw value because the tools reduced real friction in daily work. What has surprised us surprised us thus far is that the first party tools from the mobile platform vendors haven’t been the preferred tools out of the gate.

A Culture of Sharing Wins and Failures #

We opened a Slack channel to capture real use cases. Engineers shared examples where AI produced clean boilerplate or clarified complex logic. They also posted failures where the tools returned convincing but incorrect answers. Both kinds of stories built judgment. Over time, the team learned when to trust AI and when to challenge it.

The Hackathon: Clear Commitment #

We ran an internal AI hackathon early in the process. Participation was strong across the organization. Teams built automated crash analysis tools, smarter review bots, and workflow automation. Several projects shipped and everyone learned. This showed that AI was not a side project. It was part of how we improve engineering work.

Beyond Coding: AI Across the Lifecycle #

Engineers do more than write code. They write tests, design documents, user stories, and help content. AI now assists across that entire workflow.

We see engineers use AI to:

• Draft design documents and pressure test their own logic
• Write fuller test plans
• Review PRDs and surface gaps
• Create user stories from rough inputs
• Improve documentation

These gains stack. Time saved early turns into time for deeper technical work.

The Human Side: More Capability and More Joy #

We see a clear shift. Engineers learn unfamiliar parts of the codebase faster. They pick up new languages faster. Documentation and test writing move with less friction. The tools expand what someone can take on.

The work feels better. Tools evolve fast and the speed can feel overwhelming, but baseline productivity is higher. Engineers report that they are more self-sufficient.

Managers see fewer blocked engineers. Junior engineers lean on senior engineers for context instead of early unblocking. AI covers the exploratory steps so engineers focus on decisions that require judgment.

Context still matters. Judgment still matters. But the range of what an engineer can do independently is wider than before.

Tooling Is Moving Fast #

The tool landscape keeps shifting. A workflow that looked strong six months ago might be replaced by something better. Teams continue testing new capabilities as they become available. The pattern is clear: keep experimenting. Stay close to the tools and close to the engineers using them.

What Comes Next: Faster Iteration and Closer Collaboration #

We’re exploring how faster inputs from AI change our planning approach. AI helps teams draft and revise PRDs and designs quickly. This creates opportunities for product, UX, CX, and engineering to work in the same cycle instead of in handoffs.

The opportunity is tighter iteration loops. Faster documents. Faster alignment. Faster decisions. This reduces the gap between idea, design, and implementation.

The outcome we want is simple: clearer requirements earlier, better designs earlier, cleaner execution downstream.

Lessons Learned #

Here’s what worked for us:

• Set clear quality standards and create space for experimentation
• Address job anxiety with facts
• Invest in tools that match real workflows
• Create forums for sharing successes and failures
• Use focused events to accelerate learning
• Look beyond coding
• Adjust as tools evolve

The Real Win #

We’re a global engineering team building software that field technicians rely on daily. The work is demanding. The stakes are high.

AI has not made the work easy. It has made the team faster, more curious, and better equipped to focus on problems that require judgment.

That’s the change we wanted to see.