Alex Cornell’s 3 Meta Muse Workflows for Shopping, Hobbies, and Archives

Welcome back to How I AI! I’m Claire Vo, and I’m on a mission to help you build better with our new AI tools. I’ve used just about every AI agent out there, but I have to tell you, my experience with Meta's Muse delivered the fastest time to pure delight I’ve had in a very long time. It’s zippy, personal, and the design is just so thoughtful. So I was thrilled to sit down with the person who helped shape that experience: Alex Cornell, the Design Director for Muse at Meta.

Alex has been working in software for a long time, but he shared that his mom had never really used anything he’d built before Muse. That says so much about what they’ve achieved. They’ve created an agent that feels less like a sterile productivity tool and more like a genuinely helpful, personal entity. It’s designed around the patterns of messaging, not traditional software. It uses progressive disclosure to show you it’s working without overwhelming you with technical jargon. And it’s full of delightful little details, like the personalized avatars that type on tiny laptops.

In this episode, Alex walked me through how he uses his own Muse agent, a cactus named Marlowe, for some of the most creative and personal workflows I’ve seen. He showed me how he turns his scattered digital history—from Instagram saves to a 20-year-old blog—into new, useful, and often hilarious artifacts. These aren't just abstract ideas; they are concrete examples of how an agent can integrate into your life to help you with everything from shopping to improving your golf game.

We’re going to get into three of my favorite workflows from our conversation: creating a personalized shopping catalog, supercharging hobby improvement with data analysis, and resurrecting a long-lost blog as a daily podcast.

Workflow 1: Turning Instagram Saves into a Personalized "SkyMall" Catalog

One of the first things that completely captured my imagination was how Alex uses Muse to wrangle his Instagram saves. Like many of us, he’d been saving posts for years—tips, products, ideas—but that data was just sitting there, hard to access and impossible to reason with. This workflow turns that digital graveyard into a source of daily fun and inspiration.

The Problem It Solves

Your Instagram "Saved" collection is a rich, unstructured dataset of your personal interests. But the app itself offers few tools to organize or make sense of it. This workflow extracts that latent value and presents it in a completely new, engaging way.

Step-by-Step Breakdown

• Connect to Instagram: First, Alex gave his Muse agent, Marlowe, access to his Instagram account. This is the crucial first step that allows the agent to see all the content he has saved over the years.
• Organize the Chaos: His initial prompt was simple. He asked Marlowe to look at everything he’d saved on Instagram and organize it.

"I had Marlo look at everything I'd saved on Instagram and just organize them. So put all of my saves into folders and categories. Useful baseline."

• Generate Shopping Ideas: With the data organized, he took the next logical step and asked for personalized recommendations.

"Okay, look at everything I've saved, and maybe send me some suggestions for shopping ideas, like gifts and that kind of thing."

• Add a Creative Layer: This is where it gets brilliant. Instead of just a boring list, Alex prompted Marlowe to re-imagine the output with a very specific, nostalgic aesthetic.

He asked Muse to render the suggestions "in the kind of the vernacular of the SkyMall/Sharper Image magazine." The result is what he calls "Jollybot SkyMall," a daily-refreshed feed of products he’s likely interested in, complete with the wonderfully kitschy design of a '90s in-flight catalog.

Tools and Results

• Tools: Meta's Muse with its Instagram connector.
• Outcome: A delightful, personalized, and visually engaging shopping catalog that's actually useful. It's a fantastic example of taking mundane data and, with a simple but creative prompt, transforming it into something truly personal and fun. I do something similar by having my agent look for sales on items I'm interested in!

Workflow 2: Supercharging Hobbies with Data Analysis

Alex is a data-driven guy, and he applies that mindset to his hobbies. He uses several apps and devices to track his performance in golf and chess. But raw data is just numbers. The real power comes from using Muse to analyze that data and turn it into actionable insights.

Part A: Golf Swing and Round Analysis

Like many golfers, Alex tracks every round and every shot. He even uses machines to analyze his swing during practice. Muse acts as his personal data analyst and coach.

The Process:

• Feed the Data: Alex provides Muse with data from his golf activities. This includes round scores from a standard golf app and detailed shot-by-shot data captured by his watch.
• Automate a Recap: He has a recurring task for Muse. After he plays, it automatically generates a "round recap."
• Visualize and Analyze: Muse doesn't just list stats. It creates rich visual dashboards. Alex showed me a great example where a simple graph clearly illustrated his bad front nine versus his good back nine. The recap also provides specific insights, like how he could have turned an 84 into a 79.

"It's taking data that was pretty much very plain and monolithic and not really showing me more than the number and turning it into these different visuals that tell much more of a story."

• Drill into Practice: He does the same for his practice sessions, feeding data from his swing analysis machines into Muse for a detailed breakdown of what to work on.

Part B: A Snarky Chess Coach

Alex applies a similar workflow to his chess game, turning his chess.com data into a personalized coaching dashboard.

The Process:

• Connect the Account: He connected Muse to his Chess.com account.
• Define the Dashboard: He prompted Muse with his goal in a chat.

"I have a Chess account. I would love to see kind of a mixture of analysis, but also, like, trends over time... give me a review of the most recent game."

• Get Coached: The result is a dashboard that tracks his performance over time and breaks down his recent games. What's especially fun is that the agent developed a personality on its own. Alex described it as a "snarky chess coach" that pokes at him a bit, with notes like "bishop b5 collapse."

Workflow 3: Reviving a 20-Year-Old Blog as a Podcast

This might be my favorite workflow of all, because it speaks to the idea of recovering parts of our digital selves that we thought were lost to time. Alex had a blog 20 years ago, and he used Marlowe to not only find it but bring it back to life in a completely new format.

The Problem It Solves

So many of us have creative work scattered across older blogging platforms—Blogger, Tumblr, LiveJournal. This content is often hard to find and even harder to consume. This workflow resurrects that content and makes it accessible again.

Step-by-Step Breakdown

• The Treasure Hunt: Alex sent Marlowe on a mission to find his old blog from two decades ago. This wasn't a simple search; the agent had to be resourceful.

"I had Marlowe go and find it, WayBack Machine, going to Blogger, 'cause it moved between Tumblr... I moved it all around, you know?"

• Consolidate the Archive: Once it found all the posts, the first task was to gather them into a single, stable format. Muse compiled everything into one PDF.
• Transform the Format: This is the creative pivot. Reading a giant PDF isn't that enjoyable. Alex decided he'd rather listen to it. He prompted Marlowe to create a personal podcast from the blog archive.

"The way that I like to read it now, just again for memory, is via a podcast, which is sent to me every morning, and it is just basically a random reading."

Tools and Results

• Tools: Meta's Muse, its web browsing capabilities including the Wayback Machine, and its text-to-speech functionality.
• Outcome: Alex now gets a daily podcast episode featuring a random post read by a synthetic voice. He described the hilarious experience of hearing his 20-year-old self wonder about things like, "Is it weird to anyone else that 90% of people wear the exact same color pants?" It's a powerful and funny way to reconnect with a past version of yourself.

From a Cactus to a SkyMall

What I love about these workflows is how they all stem from a single, core idea: using an AI agent to transform your personal, often-forgotten data into new forms that are useful, insightful, or just plain delightful. Alex isn't just executing tasks; he's collaborating with his agent to create things that wouldn't exist otherwise.

These examples show the power of designing an agent that people want to talk to. By focusing on a messaging-based UI, providing clear-but-unobtrusive status updates, and allowing for deep personalization, the Muse team has created a tool that feels more like a partner than a machine.

I hope this gives you some ideas for your own projects. What scattered data do you have in your life? An old photo library? A messy notes app? A list of saved articles? Maybe it’s time to ask an agent to help you turn it into your own version of a SkyMall catalog or a personal podcast.

Thank you to our sponsors

Thank you to our sponsor, WorkOS!

• WorkOS gives you drop-in APIs for enterprise features so your app can become enterprise-ready and scale up market faster. Think of it like Stripe for enterprise features. OpenAI, Perplexity, and Cursor are already using WorkOS to move faster and meet enterprise demands. Join them and hundreds of other industry leaders at workos.com. Start building today.