As a seasoned Product Manager (PM), it’s either evolve or become irrelevant. Luckily for me I’ve had many opportunities to evolve, especially in the last year. In particular, I get to use AI tools to help me be more efficient.

Case in point: I recently went a bit meta and used three AI services to build my AI-based web application for musicians. The AI services I used are: ChatPRD for creating my Product Requirements doc (PRD), v0 for prototyping, and Cursor for building the app.

And they worked great!

This first article covers how I used ChatPRD. The next two articles cover how I used v0 and Cursor respectively. At the end of the last article I link to the world-class app I built.

What I love about ChatPRD is that I just gave it a brain dump of my idea and it steered the process from there. PMs don’t always have green fields ahead of them — we often have a lot of the details in place and want to improve a particular aspect of a product with a particular goal. In this case we were starting from scratch so ChatPRD asked me to “share a bit more about what [I’m] thinking.”

I gave it the following:

• One of the ways musicians make money is for companies to find the musician’s music online and use it in a production, like a video, online presentation, or physical installment. However, there is SO much music available it’s hard for musicians to ensure that their music gets found.
• One of the tasks that musicians MUST do is “tag” their music so it can be found with online music search engines. That means adding an evocative description, multiple keywords to the track’s metadata, etc.
• It’s very time consuming and laborious to come up with the right description and keywords and then add all this metadata to the audio file.

I want to call the website TrackTagging.com. But don’t steal this idea, this is my retirement we’re talking about! 🙂

After telling me I had a “fascinating concept” ChatPRD systematically went through multiple steps to nail down what it needed before creating the PRD. For example, it prompted me to consider several challenges, such as: ensuring the accuracy of the AI and training the model with enough diverse musical pieces; and building trust with musicians to encourage adoption.

Unprompted, ChatPRD then proceeded to give me a small blurb on the competitive landscape and where there may be a gap in the market that I could fill. I asked how I might go about doing a competitor analysis for this unrealized product. ChatPRD quickly gave me six clever ideas for testing out competition with companies that I didn’t even know existed.

Finally, ChatPRD asked me to provide some clear goals for the user, goals for my business, a high-level user story, basic functional requirements, and basic success metrics. In addition to my user and business goals, I provided:

Functional Requirements:

1. Listen to the track and create an accurate and evocative description, and also add relevant keywords
2. Enable the musician to edit the metadata as needed

User Experience:

1. It should be easy for the musician to upload the track from their computer.
2. The UI should be minimalist but attractive
3. The display of the metadata should be attractive and easy to edit

Success Metrics:

1. Musicians are pleased with the results, measured by how much editing they do to the metadata that’s returned
2. Musicians return regularly to submit tracks

That’s it! ChatPRD took it from there. Within a minute it came up with a glorious PRD that absolutely nailed what I’m trying to build. Check it out:

You’ll see that it created, among other things:

• Expanded goals that consider several vantage points and include tangible metrics
• Multiple user stories for my main persona (musicians) as well as a secondary user persona (producers submitting on behalf of musicians)
• The funnel for musicians to find TrackTagging.com as well as detailed steps within the core experience
• High-level UI/UX recommendations based on modern web apps
• And one of my favorite parts of the PRD: a three paragraph narrative of how a musician named Jesse struggles with tagging their tracks and how they use the service to get their music noticed. Nice!

As I said, glorious.

ChatPRD went above and beyond by sussing out plenty of detail that I would not have had the bandwidth to explore fully. The various elements of the PRD were concise and well-worded so I had no trouble understanding its recommendations. Most of what ChatPRD suggested was right on the money, and anything that was off I easily molded into shape.

I could not have asked for more.

In the next article, Building a Working Prototype with AI in Just One Minute, I cover how I use my new PRD with v0 to create a working prototype.