The Pros and Pitfalls of AI Product Development - Part 1

Building a Better App for Personal Shoppers

This is the first in a series of posts cataloging my journey into AI-augmented product design and development—using AI to help me design and build a better tool for personal shoppers.

Seizing Opportunities In A New Reality

In the last few years, AI has turned the tech industry on its head. What is asked of product managers and their teams, and what companies are looking for in product leaders, has become intensely focused on AI. AI skills, experience in AI-driven or augmented product development, etc., have become the new baseline for these roles. In fact, without demonstrable AI credentials and experience on your resume, it has become almost impossible to get a hiring manager to even glance at your resume.

As a product leader with over a decade of experience leading zero-to-one product development teams and bringing award-winning products to market, I have been considering how I can both bring these skills into my repertoire and effectively demonstrate them.

So, when I recently had a chance to gain firsthand experience with one of the leading grocery fulfillment applications used by personal shoppers, I saw an opportunity. Just a few minutes with the tool turned up mystery-meat icons, inconsistent navigational patterns, and loading screens for small bits of information that completely locked down the app. I knew I could design and build a better tool, and I would use AI to help me do it.

A Brief Note On "Enterprise" UI/UX

As a product designer, I am continually struck by the prevailing mindset that "enterprise" software should be purely utilitarian and merely "good enough" to fulfill its purpose. As a society, we dedicate incredible amounts of time and money to researching, planning, and creating "consumer" applications, having proven the ROI that good design generates in those contexts.² Yet the tools that people utilize for the majority of their workdays often receive minimal attention, despite arguably possessing equal, if not greater, ROI potential.

Why shouldn't the tools we spend the majority of our time using receive the same care and attention as our casual games?! It may not seem as "exciting," but it is crucial to recognize the importance of design in enterprise software and its impact on user experience, productivity, and the bottom line.

The personal shopping application I was using was one of the clearest and worst examples of the "good enough for enterprise" thinking I had seen in quite some time. And although that made using the app an exercise in frustration, it only made the opportunity in front of me that much clearer.

Beginning the Journey

As I considered how best to approach this project, I decided to use Claude (Anthropic) as my primary AI research and development partner, and leveraged other tools where they fit best.¹

I started by using Claude to help me conduct some initial industry research so I could better understand the space, the players, and the challenges beyond my specific experience. Once I felt I had a reasonable grasp on the space, I created three core sets of information to work from:

1. Where do the current tools (and specifically the one I am using) fail?

2. How can those failures be addressed?

3. What would the ideal personal shopping application be able to do?

After having Claude assist me with my industry research, and verifying the veracity of what we found, I took that information, my lists of issues (combined with mockups of several of the key issues mentioned), and my list of ideal features, and began the journey of turning this foundation into a full, interactively prototyped application.

These posts will focus on the process and, as the title says, "The Pros and Pitfalls of AI Product Development." They will be less about the app itself (don't worry, I'll cover that elsewhere 😉, and there will obviously be quite a bit about the app throughout the process), and instead will discuss how I leveraged AI throughout the process, and the processes I built around AI to ideate, plan, and build a real product—not just a pretty picture, or even an interactive pretty picture, but a fully functional, competitive application for personal shopping.

I look forward to having you along for the ride!

---

¹ Other tools used include Claude Design, Figma, Figma Weave, Adobe Firefly, Lovable, Cursor, Kiro, and others. I'll be sure to point out the tools I used and what I used them for throughout these posts.

² Instead of a long list of links on this topic, here is one, which includes links to three more: https://tripsixdesign.com/blog/the-roi-of-good-ux/