The Death of Search: How Shopping Will Work In The Age of AI

Imagine a web where “most things are crap,” yet they rise to the top through SEO-optimized strategies—this is the polluted digital marketplace AI agents are poised to navigate and potentially clean up. Alex Rampell and Justine Moore from A16Z explore how AI will reshape our shopping habits, questioning everything from Google’s role as a “tax on GDP” to the future of impulse buys and considered purchases. They argue that AI’s greatest impact won’t be on the emotional, spur-of-the-moment purchase or the massive, tactile buy like a house, but on the vast middle ground of researched goods. Here, agents could act as hyper-knowledgeable, objective personal shoppers, sourcing from trustworthy video reviews and forums rather than SEO-driven affiliate junk, ultimately redistributing the economic “tax” currently paid to search engines.

The discussion delves into the complex mechanics of online commerce, particularly the flawed system of last-click attribution that rewards entities like coupon extensions for “stealing” credit rather than genuine influence. As AI begins to intermediate the research and buying process, this attribution model becomes even more tangled. The conversation also highlights a crisis of trust; the open web is now dominated by low-quality, monetized content, making it hard for AI to distill honest recommendations. In response, the hosts point to models like Costco, which built immense value through curated quality and membership trust, and YouTube creators, whose in-depth video reviews offer a more reliable information source than text-based SEO farms.

Looking ahead, the most promising opportunities lie in specialized AI agents that handle specific tasks impeccably well. These could range from agents that relentlessly hunt for the lowest price and optimal cashback for commodity items with UPC codes, to vertical-specific advisors fine-tuned on expert knowledge for complex purchases like bikes or laptops. For merchants, the shift will necessitate redesigning sites to be “browsable” by AI and building infrastructure for agentic commerce. While giants like Amazon and ChatGPT will play major roles, the fragmentation of the shopping journey creates openings for new, durable companies to emerge by solving discrete pieces of the AI-commerce puzzle with superior ease and specificity.

Surprising Insights

• Affiliate marketing’s origins are in pornography: The practice of paying a commission for a referred customer, which underpins much of today’s online commerce, reportedly began as a tracking solution for the adult industry.
• Costco’s business model is fundamentally anti-profit on products: The company views high gross margins on individual items as degrading the value of its membership, so it actively works to keep product profits minimal, making almost all its net income from membership fees alone.
• Google may already be losing search volume, just not commercial queries: The hosts observe that while people are using AI like ChatGPT for informational searches (e.g., “who won the Oscar in 1977”), commercial intent searches on Google remain robust for now, protecting its core revenue.
• The “best” SKU for Amazon is advertising: Amazon’s highest-margin product isn’t a physical good you can buy; it’s the ad space on its own site that directs you away from Amazon, highlighting how its incentives can misalign with consumer goals.
• Teenage girls are early predictors of AI commerce behavior: A viral trend saw young women using ChatGPT to identify clothing and accessories from celebrity photos and find affordable dupes, signaling early adoption of AI for discovery and price-sensitive research.

Practical Takeaways

• For trustworthy product research, seek out unsponsored video reviews on platforms like YouTube. These often provide more nuanced, honest comparisons than SEO-optimized “top 10” listicles riddled with affiliate links.
• Use price-tracking tools and set alerts for specific items you know you’ll repurchase. Services like CamelCamelCamel demonstrate the consumer desire for price optimization, a behavior AI agents will eventually automate fully.
• Be highly skeptical of “last-click” attribution in your own marketing analyses. Understand that many touchpoints influence a purchase, and avoid over-rewarding channels that simply capture the final click without providing real value.
• When building or evaluating a direct-to-consumer brand, consider the durability of demand. Many trendy DTC companies falter because they don’t control manufacturing and face immediate competition from copycats; recurring revenue or a strong community can provide more defensibility.
• As a merchant, start thinking about how to make your website and product information easily parsable by AI agents. This includes clear data structure, detailed specs, and perhaps even creating AI-friendly content channels, as the browsing customer of the future may be non-human.