Building ChatPRD and 🌶️ Takes on the Future of Product Management in the AI-Era w/ Claire Vo

Building ChatPRD and 🌶️ Takes on the Future of Product Management in the AI-Era w/ Claire Vo

Building ChatPRD and 🌶️ Takes on the Future of Product Management in AI-Era w/ Claire Vo

In this week’s episode of "Unsolicited Feedback," we're excited to welcome back Claire Vo, CPO of LaunchDarkly and the founder of ChatPRD. As a repeat guest, Claire delves into the intricacies of building her AI tool, ChatPRD, and shares her incisive perspectives on the evolving roles in product management as AI reshapes the industry in 2024 and beyond.

Building ChatPRD

🚀 Claire Built ChatPRD Over Thanksgiving Break, By Herself What used to take a team of engineers six months, can now be accomplished in a few days by a single person. Claire is living proof of this. Using a stack of modern tools and APIs, she demonstrates the accelerated pace at which today’s product developers can bring AI innovations to market, effectively democratizing the ability to innovate.

🛠️ The Minimalist Stack That's Powering ChatPRD Claire utilized a streamlined yet powerful set of tools to develop ChatPRD:

• OpenAI APIs: The brain behind the AI interactions.

• Stripe: Keeps the cash flowing with seamless payment processes.

• Vercel: Used for deployment and scaling of the web application.

• Russell’s AI Package: Provides the foundation for the chat interface, out of the box.

🔥 Real Talk: The True Challenges of Creating AI-Driven Products Claire clarifies that while the business app and UI around the AI are pivotal, the real challenge lies elsewhere. Initially, setting up the business model and infrastructure for ChatPRD was straightforward. Once completed, she initially focused on tuning the AI to reliably deliver useful results. Now, she’s focused on developing a multimodal interface that shifts seamlessly between chat and structured data formats. These tasks require far more attention than simply standing up the business and are where the interesting problems to solve exist beyond just writing lines of code.

"The business model forward web app was pretty easy… Actually getting an assistant instead of instructions instead of functions that works in a relatively predictable way is a lot of work. It's what I did on my own over several months to tune it, to be as good as it is right now. A non-deterministic system is non-deterministic. It's hard to pin it down to get consistent and reliable results.

Then, where I actually spend most of my time is building differentiating features beyond the chat. Chatting is a solved problem, but what is not a solved problem is actually a multimodal interface with data that switches from chat to something more structured and back to chat. So that's where I've been spending most of my time: how do I create this data model and product that can allow PMs, in my case, to have a conversational interface, but also have an interface with the entity that they're working on." - Claire

Claire’s Spicy Takes:

Most people think being a great PM means saying NO. I think it's about saying YES" - Claire Vo

Claire challenges the traditional role of product managers as gatekeepers who primarily say "no" to manage scope and resources. Instead, she advocates for a paradigm shift towards saying "yes," which she argues can unlock greater creativity and innovation within teams.

Claire discusses how the typical restrictive PM role focused on minimizing risk can benefit from embracing a "yes" mindset, which is not just about being agreeable but about strategic thinking and exploring the full potential of ideas. This approach encourages PMs to look beyond current limitations, fostering a culture where innovation thrives and teams are empowered to achieve more impactful outcomes.

💻 AI Lets Product Managers Get Closer to the Code Than Ever Before, And They Should Take Full Advantage of this Benefit Claire encourages product managers to engage directly with coding to truly understand and influence the development process. She shares her personal experience with ChatPRD, where being involved in coding allowed her to rapidly iterate on product features. This hands-on approach not only speeds up development but also enhances the overall quality and relevance of the product.

📊 The Problem with BI Tools: It Leads to Shallow Analysis While we see PMs getting closer to the code, we see them getting further from the data. Business intelligence tools democratize access, but also lead to superficial analysis. They may make it too easy to generate reports without deep engagement.

We support high-friction data analysis. Deeper engagement with data leads to more substantial insights than surface-level analysis. There are real advantages of writing SQL inquiries yourself, and understanding how data charts are built, rather than just looking at them. Truly insightful analysis requires a deeper, more hands-on approach to data, beyond what most BI tools offer.

🤖 We Think AI May Lead to More Junior Hires, Not Less Fareed posed the question on a lot of our minds, “Will AI propel companies to only hire experienced talent.” Challenging common perceptions, Claire suggests that AI tools like ChatPRD could lead to an increase in hiring junior talent. AI's ability to enhance productivity and decision-making could make less experienced personnel more effective, thereby shifting hiring trends to favor these new talent who can deliver greater value than ever before thanks to AI assistance.

Next up, we’re discussing Loom and Linear! See you Thursday!