PodBrowser
a16z Crypto

The end of ads? AI agents are about to change how we buy

Monday, 27 April 2026 · 3 min read · Listen to the episode ↗

The podcast discusses the impact of AI agents on commerce, suggesting they may replace traditional apps by automating purchasing processes and enhancing consumer interactions. It highlights the emergence of an agent-native internet that could diminish the role of advertising, as user behavior shifts towards AI-driven transactions. Additionally, the conversation addresses the potential transformation of payment systems, considering the integration of stablecoins to improve transaction efficiency and reduce costs associated with traditional credit card systems.

The podcast explores the transformative potential of AI agents in automating commerce processes, questioning the future necessity of traditional apps. Sam Ragsdell categorizes AI agents into two types: conversational commerce, which enhances consumer experiences, and agents that can spend money on behalf of users, though they currently struggle with complex tasks. The discussion highlights a shift towards an agent-native internet, where traditional search methods may become obsolete, and agents could independently purchase goods and services, fundamentally altering online commerce dynamics.

Eddie Lazarum notes that recent improvements in AI models allow agents to understand user intentions through conversation and memory. Challenges in agentic commerce include establishing authority and identity for agents and managing payment processes. The podcast emphasizes advancements in preference modeling and intent formation, which streamline user interactions from intent to fulfillment. AI agents are now capable of performing tasks comparable to average human computer usage, often more efficiently and at a lower cost, raising questions about the future of applications and user interfaces.

The concept of "headless merchants" is introduced, where AI agents manage transactions, transforming various types of commerce and democratizing access to development tools for B2B applications. Developers seek flexible access to resources, emphasizing the need for per-usage models without lengthy onboarding. Traditional businesses often rely on friction to foster customer loyalty, but as productivity rises, the costs associated with this friction become more apparent.

The conversation also addresses the changing economic contract of the web, moving away from content publishing and advertising towards a model prioritizing efficiency and user experience. The rise of internet advertising revenue contrasts with the evolving role of agents in information retrieval and task execution. Users primarily engage with tools like ChatGPT for search-like queries rather than as agents, indicating a significant shift in user behavior.

Concerns are raised about whether AI agents might succumb to distractions similar to human users, emphasizing the importance of understanding whose interests the agents serve. The potential for advertising to re-enter agent interactions is explored, including how companies could influence agent responses through financial agreements. The challenges faced by foundation model providers, particularly OpenAI's ChatGPT, are highlighted, noting that while advertising could provide a monetization solution, major models currently avoid ads to maintain neutrality.

The role of agents in commerce raises questions about how they will navigate and differentiate between merchants. Predictions indicate that as agents take over the buying process, implicit advertising methods may emerge, leading to discussions about the adaptability of traditional payment systems. The effectiveness of credit cards in e-commerce is acknowledged, particularly due to consumer protections, while the potential benefits of stablecoins for online transactions are considered.

The conversation centers on the evolving landscape of payment systems, particularly the potential shift from traditional credit card transactions to innovative solutions like stablecoins and AI agents. The inefficiencies of credit card transactions are discussed, particularly the multiple parties involved that lead to unnecessary fees. There is a proposal for credit card companies to mint stablecoins, creating a hybrid payment model that combines the benefits of both systems.

The potential for agentic transactions is explored, emphasizing how AI agents can streamline various tasks, enhancing the shopping experience. The current credit card system's limitations, such as restricting transaction processing to 80 characters, highlight the need for improved data presentation. Data management is crucial, with companies leveraging AI to enhance clarity and usability. The introduction of stablecoin payments could address existing financial issues, paving the way for more efficient transactions.

The conversation concludes with a consensus on the promise of innovations in payment systems and commerce, suggesting a revisit of predictions in five years.

This summary was generated from the episode transcript and can contain mistakes.