Augmenting Software Development with AI
For years, we’ve talked about the potential of AI in software development. Now, it feels like that future is finally arriving, and we’re actively trying to figure out what it means for us. At my current work, I've embarked on a journey to explore how AI tools can augment our mobile development workflows, and I wanted to share some of my early findings.
As a developer on the mobile team, I’ve been part of a small group piloting several AI platforms: Copilot, Cursor, and Claude Code. The goal isn’t to replace developers (far from it!), but to identify tools that can streamline repetitive tasks, accelerate problem-solving, and ultimately allow us to focus on the more creative and strategic aspects of our work.
Why Now? And What Am I Looking For?
The recent advancements in large language models (LLMs) have been genuinely impressive. These tools aren’t just about auto-completion anymore; they can understand context, generate complex code structures, and even help with debugging. I’m specifically looking for tools that can:
Reduce boilerplate code: Mobile development often involves a lot of repetitive tasks. Can AI help us generate common components and structures more quickly?
Accelerate problem-solving: Can AI help us identify potential bugs, suggest solutions to complex problems, and even generate test cases?
Improve code quality: Can AI help us identify potential vulnerabilities, suggest improvements to our code style, and even generate documentation?
Integrate seamlessly into our workflow: Can these tools work with our existing IDEs and version control systems without causing friction?
Early Observations: A Mixed Bag (But Promising)
I’ve been experimenting with these tools for a few weeks now, and the results have been mixed.
Copilot: Excellent for auto-completion and generating simple code snippets, but sometimes struggles with more complex tasks.
Cursor: A powerful IDE that integrates AI features seamlessly, but can be resource-intensive.
Claude Code: Prioritizing security with a direct API connection to Anthropic. I’m also exploring how to integrate these tools with our existing CI/CD pipeline to automate code quality checks.
The Power of Connection: Leveraging Machine Control Protocol (MCP)
One particularly exciting development has been leveraging the Machine Control Protocol (MCP) to connect these AI models to our existing tools like GitHub, Jira, Slack, and more. This allows us to automate tasks like creating pull requests, updating bug reports, and posting notifications, significantly streamlining our workflow.
Ultimately, my goal is to leverage AI responsibly and improve the tools available to our development team. I believe that AI has the potential to augment our capabilities and allow us to focus on what we do best: building great mobile apps.
