How I Built a PM Mock Interview App with no coding skills
From a Product Manager to an AI Powered Prototype Manager
Hi, My name is Keerthana and welcome to my space - No BS Product Management, your weekly guide to mastering the product management journey—from breaking into the role to crushing it as a PM. Every week, I share straightforward advice, actionable tips, and proven strategies to navigate tough situations, win over stakeholders, and elevate your PM skills.
If you’re ready to stop guessing and start thriving in the world of product management, hit that subscribe button now! 🚀
As Product Managers, we constantly identify problems and think of solutions. But how often do we take those solutions from ideation to execution? For many PMs, the technical barrier stops them in their tracks. I’ve been there too. Without a solid programming background (though I understand system design), the idea of building an app seemed intimidating.
But recently, I challenged myself to build a solution to a recurring problem I’ve seen in the PM community: interview anxiety. Many aspiring PMs struggle to practice for interviews, even mock sessions with peers, fearing judgment or not receiving actionable feedback.
That’s when I thought of creating a safe space for PMs to practice interviews. An AI-powered app that would:
Provide a comprehensive question bank covering key PM skills.
Allow users to record their answers for realistic practice.
Offer specific, actionable feedback to help them improve.
I had this vision, and to bring it to life, I turned to Bolt.new—a platform that lets non-programmers like me build functional apps without losing our sanity. After some trial and error (and a few “what the hell is this error?” moments), I built a fully functional PM Interview App. It took me about 10-12 hours in total, spread over a few days—partly because I kept tweaking ideas, and partly because Bolt had a habit of throwing unexpected errors every time I changed a component.
Here’s how I did it and how you can too.
What’s the App About?
I wanted the app to be a place where users can practice mocks without anxiety and get actual contextual feedback.
Comprehensive Question Bank: It includes categories like Product Sense, Strategy, Estimations, and Stakeholder Management to ensure holistic interview preparation.
Voice Recording for Realism: Users can record their answers, simulating the actual interview environment.
Actionable AI Feedback: The app provides detailed, specific suggestions to improve performance.
With this setup, candidates can build confidence and refine their answers before moving to peer mocks or real interviews. It’s fully functional and can be used
How I Built the App Using Bolt.new
1. Starting with the Idea
The first step was translating my idea into a prototype. Bolt makes this easy: you describe your vision, and it generates a functional skeleton of the app. I outlined the app’s purpose and key features, and Bolt quickly gave me a basic working version.
The core idea was to get a PM interview question, user would answer via voice preferably, and they would get tailored feedback as if it were a regular interview.
Once the idea was in place, I give bolt the prompt to create the app, something like “Create a mock interview app for Product Managers with questions from different categories - Product Sense, Estimations, Stakeholder management and Strategy. Allow them to choose questions from any category. They can answer through voice or typing, and then the answer is passed to open AI API to get relevant feedback”. This creates a basic app flow, which can be further tweaked by giving similar prompts.
2. Setting Up Integrations
To make the app functional, I had to have a relevant question, voice input for user, and contextual feedback for the answer provided by the user. I wanted to generate the questions using the app itself, however, I didnt want it to hit the api everytime because it would be costly. So, I pre-generated 50+ questions. I wanted the answer, however, to get contextual feedback which meant that every answer is hitting the Open AI api. Because I wanted to open it up for free to other users, I wanted to limit the tries to 3.
I had to have a question bank which I generated using prompts, to get the feedback, I had to pass it to Open AI API which and structure the prompt to give proper structured feedback. Then used netlify to deplot the app. So, all these needed have some integrations set up - Thankfully, Bolt has all these built in.
To make the app functional, I did these integrations:
OpenAI API: This powers the feedback system, generating actionable insights based on user responses.
Supabase: An open-source backend-as-a-service platform that provides a database solution, making it easy to store user responses and track activity. Supabase simplifies backend setups and offers real-time capabilities, allowing me to store and retrieve practice session data effortlessly.
Netlify: A powerful platform for deploying and hosting web apps. It automates builds, integrates seamlessly with Git repositories, and ensures fast, reliable performance through its built-in global Content Delivery Network (CDN). Using Netlify, I could share the app quickly without worrying about server management or downtime.
Bolt’s built-in integrations for these tools made setup straightforward. All I had to do was:
Configure my OpenAI API keys.
Connect Supabase for backend tracking and data storage.
Deploy the app on Netlify for a professional-grade hosting solution.
3. Iterating and Refining
The initial prototype worked well, but I wanted to enhance the user experience. Here’s how I iterated:
UI Improvements: Bolt did give me a 60% ok prototype, but there were a lot of things that I adjusted, so that the layout and design looked more intuitive.
Feedback Quality: The AI’s feedback is the app’s core feature. To improve it, I refined the prompts, ensuring the suggestions were actionable and relevant.
During this phase, I used ChatGPT extensively to tackle several aspects of the app’s development process. Here’s how it helped me:
Debugging Minor Issues: For example, when my OpenAI API key wasn’t being accepted due to formatting errors, ChatGPT provided clear explanations and solutions, which I then implemented in Bolt.
Improving Prompts: Crafting effective prompts is key to getting high-quality output from AI. ChatGPT helped me rework and refine prompts for Bolt, ensuring the system produced accurate and actionable results.
Refining Requirements: The feedback feature is the heart of the app, and I needed it to be as specific and actionable as possible. I used ChatGPT to brainstorm and define the exact requirements, which improved the clarity and usability of the feedback system.
I think my familiarity with chatgpt more than Claude maybe the reason why I used it more.
Building the app wasn’t entirely smooth sailing. Here are the key challenges I faced and how I overcame them:
1. Code Overwrites
Bolt sometimes rewrote entire files for small changes, which could disrupt functionality. For instance, while trying to adjust a minor UI element, the platform rewrote critical sections of the feedback logic, creating bugs. To avoid setbacks, I made it a habit to take regular backups and carefully review changes before deploying.
2. Debugging Loops
At times, Bolt got stuck in loops trying to fix issues. For example, when I attempted to enhance the voice recording functionality, Bolt repeatedly generated conflicting code snippets. Using console logs helped me pinpoint where the logic was failing. I also guided Bolt with alternative prompts to change its approach, breaking the loop effectively.
Another trick that is very helpful is to tell Bolt explicitly “Don’t code, just analyse and tell” - it helped me understand the logic and suggest steps to perform, instead of it going in a loop and changing things. I saw that repeatedly it would take more complex logic steps to solve easy problems instead of using a straight-forward easy way.
3. AI Understanding
A lot of times because of Bolt going around in circles trying to fix or change code, I began prefacing my queries with “Don’t code yet, just explain” - This helped me understand the code better - Sometimes, I found that it was taking complicated routes for simple changes, and I brought it back to the right direction, It was either because I was not verbose enough, or it just made some assumptions and ran with it.
Despite these hurdles, the process was incredibly intuitive for someone with no programming experience. With a combination of persistence, creative prompting, and external tools, I was able to overcome these challenges.
Key Learnings for PMs Who Want to Build Apps
Overall, at this stage I think AI is like a Junior Developer, than a Mid or Senior Developer. At many times during my experience, it was picking complicated paths to solve simple problems. I had to stop, ask and redirect.
Somethings I understood:
You Don’t Need to Be a Programmer: Tools like Bolt lower the technical barrier, allowing PMs to focus on the user problem rather than the technical implementation. However, if we don’t understand basic system design or programming logic, it becomes tough to debug after a while.
Leverage AI for Iteration: ChatGPT and Bolt can work together beautifully. Use one for refining ideas and the other for execution.
Backup Your Work: AI platforms can be unpredictable, so regular backups are a lifesaver.
Prioritize the User Experience: Even in a prototype, usability and clarity matter. Small details, like actionable feedback, can make a big difference.
Why PMs Should Try Prototyping
Prototyping isn’t just for engineers or designers. As PMs, we’re closest to the problem and its potential solution. Building prototypes allows us to:
Test ideas quickly.
Communicate better with stakeholders.
Develop a deeper understanding of the product development process.
With platforms like Bolt, the barriers to entry have never been lower. Whether you’re a seasoned PM or just starting out, building your ideas has never been more accessible.
From Idea to Execution
Building this app was an eye-opening experience. It showed me that the biggest limitation is often our own hesitation. With tools like Bolt and a bit of perseverance, anyone can bring their ideas to life.
If you’re a PM with an idea you’ve been sitting on, why not give it a try? Bolt.new might just be the launchpad you need.
Let’s democratize app-building and empower PMs to create impactful solutions. 🚀
Get Access to the App by Subscribing
I will share free early access to the app to 50 of my subscribers, who I will pick randomly. Please subscribe below to get access.