Solving Silly Problems with AI: An Alexa-to-Walmart Grocery Story

Sometimes the simplest ideas turn out to be surprisingly hard. Here's how I used Claude Code to bridge Alexa shopping lists with Walmart+ grocery — and what it says about the power of AI tools for everyday problem-solving.

The idea that should have been simple

It’s always interesting to me when I have an idea for something, and then find out that it’s infinitely more difficult than it should be.

Case in point: we have an Amazon Echo device in our kitchen. We’re pretty used to shouting to Alexa to add things to the shopping list when we use the last of something, and we’ve trained the kids to do it too. We also use Walmart+ to keep a running list of grocery items that aren’t a Costco purchase.

Silly me — I thought there would be an easy way to hook up Walmart grocery to Alexa. Maybe instead of adding to a shopping list, we just add it to Walmart grocery directly.

Nope. Not a thing.

Building the bridge with Claude Code

Fast forward, and with the help of Claude Code — I now have a script that runs on an always-on computer on my network. Every half hour, it:

Pulls up the Alexa shopping list
Logs into my Walmart grocery account
Searches for each item
Ignores promos and other distractions
Chooses the top result, with preference for items we’ve purchased before
Clears the Alexa shopping list

I also added a button on my phone to ping the server directly when I want a more immediate sync.

Handling the real-world messiness

Of course, nothing runs perfectly forever. I added the ability for the server to send me iPhone notifications when something goes wrong — typically when one of the services logs out or hits a captcha challenge — using ntfy push notification services.

Then I noticed Walmart.com was increasingly throwing bot challenges — specifically a Perimeter X challenge that asks you to click and hold a button. I told Claude about the new situation, and it built out code to automate that process (8.7-second hold time when it last ran) and proceed forward. If it can’t solve automatically, it sends me a notification.

It’s also now randomizing waits and hold times, adding a few delays into the scripting so that it looks more human. The kind of resilience engineering that would have taken me hours to research and implement — done in minutes.

The bigger takeaway

All that to say — it’s insane what kinds of silly little problems and minor annoyances we can solve through the use of these amazing AI tools and a little ingenuity.

This wasn’t a mission-critical business process. It wasn’t a million-dollar integration. It was a family grocery list. But the fact that I could go from “this should exist” to a fully working, self-healing automation — complete with error notifications and bot-challenge handling — in a fraction of the time it would have taken even a few years ago? That’s the real story.

The barrier between “I wish this existed” and “I built it” has never been lower. Whether it’s connecting your smart home to your grocery service or building agentic workflows that transform your lifecycle marketing, the tools are here. The only question is what problem you want to solve next.

What projects are you all working on?