Getting Development Teams to Actually Use AI
There’s a lot of discussion right now about AI transforming software development. The tools are impressive, but the real challenge isn’t the technology — it’s adoption.
In my experience, the hardest part of introducing AI into a development organization is the human side.
Developers are naturally skeptical of tools that promise to generate code, automate workflows, or change established ways of working. Good engineers have built their careers on precision and control. Handing part of that process to an AI assistant can feel uncomfortable at first.
I’ve worked through this transition with my team. We introduced Builder.io to convert Figma designs into Angular components, significantly reducing the manual front-end work required to translate design into production code. At the same time, we adopted Cursor to support AI-assisted development across the codebase.
The potential efficiency gains are real:
▪️Faster translation from design to working UI ▪️Reduced boilerplate coding ▪️Accelerated prototyping and iteration ▪️Improved developer productivity on repetitive tasks
But tools alone don’t create those outcomes.
The key lessons have been about people and process:
1️⃣ Adoption must be led, not assumed. Developers need time to experiment and build trust in new tools. Mandates rarely work.
2️⃣ Early wins matter. Demonstrating how AI can eliminate tedious work quickly changes the conversation.
3️⃣ Guardrails are essential. AI-assisted development requires oversight. Code reviews, architectural standards, and security practices become even more important when code is generated or assisted by AI.
4️⃣ AI should augment engineers, not replace judgment. The best outcomes come when experienced developers use these tools to extend their capabilities rather than outsource thinking.
When implemented thoughtfully, AI can dramatically increase engineering leverage. But the transformation isn’t purely technical — it’s organizational.
Getting the culture right is what ultimately determines whether these tools deliver real value.