Isenberg points out Amazon earns tens of billions from ads and airlines earn heavily from seat upgrades, bags and checkout credit-card offers. An agent still completes the sale but skips sponsored products and upsells, taking away the highest-margin parts of the purchase journey that finance those businesses.
Isenberg expects a general agent identity and payments layer to be won by Visa, MasterCard or Cloudflare. But he says every industry has messy systems, like contractor booking or vet scheduling, so someone can build the trusted connection between agents and businesses in one vertical and own it.
Isenberg describes a three-rung trust ladder for agents: research, prepare and pay. Most companies tolerate research because it resembles search, but many block all three rungs even though the real financial risk only appears on the pay rung, so they should gate payment at checkout instead.
Isenberg contrasts human brand loyalty with agent loyalty, saying agents return to whatever worked last time. If your site gets the order and delivery right, the agent remembers; if checkout gets stuck, it quietly goes elsewhere and you may never know the customer was lost.
Walmart said Meta's Muse agent failing on its site was an accident, and Isenberg notes many blocks come from security systems built to stop scrapers that cannot distinguish a scraper from a personal assistant. He recommends auditing old protection and deciding on purpose whether to block agents.
Greg Isenberg defines an AI agent as a virtual customer with a credit card and no eyes. It reads product details, prices, shipping costs and return policies, but ignores homepage videos, banners, pop-ups and countdown timers, so everything designed for human eyes becomes mostly invisible to the buyer.
In his stage-eight prediction, Isenberg says companies will pay to be the first thing an agent recommends, and notes ChatGPT has tested ads inside AI answers. Firms block agents because they skip ads, but he expects advertising to evolve rather than disappear, just like earlier formats.
