Autonomous AI agents that shop and transact on behalf of users could overturn the current online advertising economy, argues a16z Crypto. Sam Ragsdale, co-founder of Merit Systems, says “agentic commerce” will change how products are discovered and bought, reducing reliance on ads that monetize fleeting human attention.
Ragsdale characterizes the internet’s commercial model from roughly 1997 to 2024 as one built on distraction: platforms earn by capturing users’ partial attention and selling that exposure to advertisers. Large language models and agentic systems, he notes, don’t operate by the same attention-driven dynamics, and that undermines the basis for many existing ad formats. With the global online ad market estimated at about $291 billion for 2025 by Mordor Intelligence, the potential disruption is substantial if agents begin making purchases autonomously.
Some AI services are already embedding commerce into conversations. Features such as “Instant Checkout” in ChatGPT and Gemini let U.S. users complete purchases directly inside chat, removing the need to visit merchant sites. Ragsdale expects that as agent-driven buying scales, consumers will find better matches, merchants will see higher conversion rates, and platforms could capture fees in the neighborhood of 5–10% for facilitating transactions.
But he warns of a trade-off: many built-in checkout experiences can become walled gardens. When merchants must be approved by a platform to participate, openness and merchant choice shrink. Ragsdale contrasts closed, curated systems with open protocols that allow agents to discover and transact across a wide range of sellers without platform gatekeeping. He uses an analogy: an agent limited to pre-approved vendors is like an employee with a corporate card, whereas an agent operating with open protocols is like an entrepreneur with a bank account.
To counter closed ecosystems, Ragsdale highlights emerging standards aimed at enabling open agentic commerce, including Coinbase’s x402 protocol and the Machine Payments Protocol (MPP) developed by Tempo and Stripe. These approaches are presented as alternatives that could let agents autonomously search, compare, and pay across the open web instead of routing commerce through a single platform.
Ragsdale also frames advertising as a “clever hack” that helped create the free, open web and the enormous datasets used to train large language models. But as agents take over discovery and purchasing, the economics of online discovery and monetization could shift: traditional ad formats may decline while value migrates toward agent services and the payment and discovery protocols that enable open commerce.
If agents reliably handle search, comparison, and transactions across the open web, the incentives and revenue streams that supported advertising-based platforms may change fundamentally, reshaping how the internet is monetized and who captures value in digital commerce.