Inside Anthropic's strategy to win AI shopping from the inside out

Good evening.

First, I gotta be honest with you. I'm still flying high from attending Tuesday night's epic New York Knicks game at Madison Square Garden. (Condolences to Cleveland Cavs fans among you; Knicks fans have definitely been there.)

One of the few things I'm as passionate about as building The Aisle is my New York Knickerbockers…for better or worse.

As a teen, I was lucky to attend another magical Knicks playoff game way back in 1999 with my own father, who passed away eight years ago. I can still recall the joyous pandemonium of the arena on that late-90s night when our team won in miraculous fashion. I remember bear-hugging my dad, and every stranger within reach, as hard as my teenage body could. It remains one of my favorite childhood memories, and that's what made last night so special: I got to share the experience with my 13-year-old son, who I've unfortunately cursed with the same unhealthy obsession. For one night, though, it was all worth it.

So what’s the connection between my personal story and what you're actually here for? I'll just say there's little that beats the unscripted, spontaneous joy of a great real-world experience. Last night, for me, it was basketball. For others, it might be the thrill of discovering something they didn't know they wanted, had trouble finding, or couldn't previously afford. The best shopping experiences have always had those qualities. And my hope is that the architects of commerce-related AI don't work to replace those feelings, but amplify them.

Which brings me to Google. The company this week announced a bevy of new AI tools that aim to turn just about every Google consumer product into a commerce experience. Those paying attention to this space already know that Google has launched purchasing capabilities within Gemini and Google Search for products from retailers and e-commerce platforms like Shopify that sign on.

But with the imminent launch of a shopping hub it's calling the Universal Cart, Google is making it possible for consumers to save products found in a Gmail message from a retailer, or easily purchase something they discovered in an ad on YouTube.

Google executives are adamant that users want these tools showing up wherever they interact with Google online, and that they're aimed at replacing the tedious parts of shopping rather than the enjoyable ones.

"This is something shoppers tell us they want," Google exec Suresh Ganapathy told me this week. "When they're trying to buy, say, a couch...they're looking for some products, they find a few that they like on search, subscribe to a few emails from merchants, get an update over Gmail...this journey starts to feel very fragmented. The cart makes it easy for them to collect those items."

The cart will also track price drops and deals, and support products from multiple retailers. When a shopper is ready to buy, they can complete the purchase within Google for participating retailers, or have their cart transferred automatically to a retailer's own site. It’s a compelling idea.

I do wonder how aggressively retailers will integrate with these programs over time, and how well the functionality will actually work in practice. More on all of that in the coming months.

But the most important question remains unanswerable for now: how many real problems will shoppers actually feel Google is solving here? Nothing else matters.

Now on to the good stuff…

According to data from the finance startup Ramp, Anthropic recently overtook OpenAI as the top paid AI service among U.S. businesses. Much of the credit goes to Claude Code, Anthropic’s AI coding tool that has exploded in popularity inside software development divisions over the past six months, including in the retail industry. (I hope to report back soon on my own experience attempting to build with Claude Code.)

Yet when it comes to the topic of AI-powered shopping, Anthropic doesn't get a lot of press attention. That’s at least partly due to the fact that unlike ChatGPT, Perplexity, and Google, Anthropic hasn't made a play to get consumers purchasing products through Claude. Instead, it's mostly been focused on helping enterprise customers—including Amazon—improve their own shopping and AI experiences. So far, the strategy seems to be working.

I recently spoke with Kate Jensen, Anthropic’s head of the Americas, about the philosophical differences in Anthropic’s approach; why Amazon’s AI shopping assistant got noticeably better when it integrated Claude models; and what she thinks personalized shopping experiences will actually look like in a year or two.