The Agentic Wall: Why the Open Web is Closing Doors on AI
AI-generated image · Bay Street Wire
OpenAI, Meta, and Google promise autonomous assistants that can manage our lives, but a growing coalition of retailers and financial institutions are proving that 'agentic' AI is fundamentally at odds with current web architecture.
The current AI gold rush is pivoting toward 'agentic' workflows—the promise that autonomous bots from OpenAI, Meta, Google, and Instinct can scour the internet to book dinners, clear inboxes, or buy shoes. But as these agents attempt to move from chat boxes to actual execution, they are hitting a hard wall of corporate resistance and legacy security architecture.
As first reported by The Verge, the dream of a seamless, agent-driven web is being dismantled by the very companies these agents need to access. It is not just a matter of technical glitches; it is a systemic rejection. Delta Air Lines is preventing agents from booking tickets, while JPMorgan Chase and Charles Schwab have confirmed they are blocking AI agents from accessing accounts, citing security concerns. Even Yelp has restricted access, telling TechCrunch that it only permits agents that pay for its data licensing program.
**Opinion:** From a practitioner's perspective, this is the inevitable collision between the 'agentic' fantasy and the reality of the monetized web. The friction isn't an accident; it's a business model.
Take the legal war between Amazon and Perplexity. Amazon previously banned Perplexity’s agentic shopping features and filed a lawsuit, arguing that third-party companies acting on behalf of customers should respect a provider's decision to participate. Perplexity countered that Amazon's hostility stems from a loss of human eyeballs on advertisements. This highlights the fundamental incompatibility: AI agents optimize for efficiency, while the modern web is optimized for attention and ad impressions. While an appeals court overturned a ruling that ordered Perplexity to stop shopping on Amazon in August, Amazon has since filed an amended complaint.
This battle has expanded to the biggest players in the field. Meta’s Muse, xAI’s Grok, and OpenAI’s Dots have all reportedly been unable to use Amazon. Amazon specifically stated that Muse’s continued access violates its site policies. In response, Elon Musk, the owner of xAI, suggested he might start a competitor to Amazon.
Beyond explicit bans, agents are triggering the web's existing immune system. Chris Canal, CEO of the AI safety evaluator EquiStamp, told The Verge that AI agents often trigger ubiquitous anti-bot safeguards designed to stop hackers and scammers. This is evidenced by reports on Reddit of Google blocking accounts using Muse for 'suspicious activity' and users experiencing checkout failures on Walmart, despite Walmart's partnership with Meta.
Some companies are attempting to bypass this friction through partnerships. Meta has collaborated with retailers including Gap, Sephora, Best Buy, and Walmart to allow Muse to shop on their sites, and Shopify has opened its storefront network to agents. However, as Canal notes to The Verge, if companies actively work to prevent AI permeation, the perceived value of the technology drops because there are simply fewer useful things the agents can actually do.
As Cloudflare provides tools for sites to implement one-click blocks on AI visitors, the 'open web' is becoming a series of gated communities. The agentic future cannot exist if the destinations remain closed.