Like many in the tech industry, I’ve heard all about the incredible things AI tools can do. While some in tech fear that AI could disrupt their way of working, I fall into the camp that believes AI will empower individuals to accomplish more than ever before. I use ChatGPT as my daily fact-finding engine, and my team has used AI tools to storyboard user journeys. Yet, my search for the holy grail—an AI tool that could build an app from product requirements—had been fruitless. There are plenty of tools that help create better wireframes, write more effective PRDs, and accelerate product creation, but none that do the actual building. That changed when I found Cursor.
My background is in product management and startups. I first saw The Social Network in high school, which led me to read Zero to One, and I was hooked. I fell in love with startups and tech because it’s a field where your name, background, and appearance don’t matter. If you can provide value to users, you can win—and I wanted to win. In high school, I began tinkering with ideas after football practice, building spreadsheets and algorithms to bring pro baseball analytics to college baseball, inspired by Moneyball. I’ve since launched and scaled products in healthcare, fintech, banking, and marketplaces. I've helped a company progress to and through an IPO. I've been on the management team of a startup that raised a Series A round. My obsession with startups remains, and I believe AI tools will allow driven individuals to move faster than ever.
With my background in fintech, I decided to test Cursor by building a simple iOS payments app. Since this was just a test of Cursor's capabilities, I wasn’t focused on the business viability of the idea. To make the test realistic, I targeted the app at kids, naming it “Toad” (inspired by the abundance of toads in my backyard). I followed a tutorial on Twitter/X (I’ve since lost the original link, but there are many if you search for “Cursor tutorial”). Below, I’ll outline the steps I took to build the app.
Before diving in, let me offer a few caveats. First, this app isn’t 100% ready for the App Store, but it's close, and Cursor can guide you through each step. Second, I’m not a developer. While I’ve completed SQL certifications and Codecademy courses and understand development mental models and frameworks, I’m far from being able to build software, either front-end or back-end. That’s where Cursor’s magic comes in.
Building with Cursor
Step One
I began with Figma, the industry-standard design tool. It’s simple to use, and its tutorial will quickly get you up to speed. My goal was to create a basic home screen for the Toad app using iOS 18 design inputs. I quickly mocked up Toad’s homepage, incorporating Oakland A’s green and gold as a tribute to their final game in Oakland. The design wasn’t fancy, nowhere near release quality, and lacked finishing touches—but it served the purpose of testing Cursor’s capabilities.
I stopped here, took a screenshot of the design, and prepared to upload it to Cursor.
Step Two
Next, I downloaded and set up both Xcode and Cursor, signing up for a Cursor subscription (about $20/month) to access its AI tool, Composer. Once everything was downloaded, I ensured both apps were running smoothly.
Step Three
I won’t go into too much detail on Xcode since the onboarding process is straightforward. Cursor also offers a helpful onboarding guide, which I recommend following. The next step was to create a new .txt file in Cursor titled "Instructions" and upload the Figma screenshot. Here’s where it clicked for me: I could explain to Cursor in plain text what I wanted to build, just as I would explain the idea to a key stakeholder or team member.
One of Cursor’s strengths is its intelligent autocomplete feature. For example, when I mentioned we were building an iOS app similar to Cash App, Cursor suggested using Stripe for payment processing, along with SwiftUI, Swift Data, and Swift Concurrency.
I continued giving Cursor more detailed instructions—describing how the app should match the screenshot, what the menu bar should contain, login requirements, and the placement of various buttons. I concluded by instructing Cursor to complete the app.
Step Four
This is where the magic happened. Composer, Cursor’s AI tool, appeared as a simple text box, and I interacted with it just as I would with a software engineer. I instructed Composer to follow the 16 lines of text I had written and use the single screenshot from Figma.
From these basic instructions, Composer generated two files (where the actual code resides), giving me the core of my app. Composer also explained its process, offering insights as if it were a senior engineer guiding a junior developer.
Step Five
Now it was time to take the files Composer created and move them into Xcode. If Composer acted as the senior engineer, Xcode was the tool where the app would be built. As a first-time iOS app creator, I found Composer’s guidance invaluable, helping me set up Xcode, organize files, and manage naming conventions.
Step Six
That’s really all it took to create a functioning iOS app front-end. From here, I iterated on the app to refine the final product. For example, I updated the entire app’s color scheme to match the Oakland A’s colors.
I also adjusted specific UI elements, added links (e.g., to iMessage), created new pages, and fine-tuned the language within the app. Along the way, I encountered errors in Composer’s code, but each time I reported the error code, Composer swiftly provided solutions.
Conclusion
After about half a day of back-and-forth, I had a fully functioning iOS app front-end running on my machine. It included a QR reader, customer support functionality, sign-up flows, error states, Stripe integration, and more. While back-end work is still required, along with legal and compliance considerations, this was a phenomenal leap forward for me. I’m someone with no prior app development experience who managed to bring an app to partial life.
My key takeaway is that the "holy grail" of turning plain text into a fully developed app still isn’t fully solved by AI. However, Cursor’s Composer is the next best thing. I agree with successful entrepreneur Nikita Bier, who says that app companies should be led by "product architects"—people who go beyond the traditional roles of product owners or managers. I believe tools like Composer will elevate the best product managers, turning them into supercharged product architects. In the near future, the role of product owner could be eliminated, with higher standards for product managers across the board. The best startups will leverage these product architects and their mastery of new tools to generate great products with fewer resources, faster than ever before.
I'd love to hear from you about what tools you're using and what you're building with them. You can reach me at linkedin.com/in/bryceam, x.com/bryceam, and mcphail . career @ gmail . com.
