CharlieKey takeaways from your favorite podcasts
← 22 October 2026The Startup Ideas Podcast

The 9 Stages of AI Agent Adoption, Explained

Greg argues that companies blocking AI shopping agents are acting from fear and protecting ad and upsell revenue, and that the fight will play out in nine stages that open startup opportunities.

Watch on YouTube ↗

Charlie's summary

  1. Blocking agents is not universal: companies fall into block, accidentally-block, and welcome camps, shaped by how they make money

    Greg lays out three camps. Amazon blocks Meta's Muse agent and Delta says it works with zero outside agents. Yelp admits agents only if they pay. Zillow and Walmart sit in the middle, where anti-scraping defenses block agents by accident because sites can't tell a scraper from a personal assistant. Meta, Shopify, Visa and Mastercard welcome agents. Greg thinks which camp a company lands in reveals how it makes money.

    ▶ 0:00
  2. An agent is “a customer with a credit card and no eyes”

    Greg's reframe: an agent doesn't see your homepage video, banner, popup or countdown timer. It reads product details, prices, shipping costs and return policy. Everything designed for human eyes is mostly invisible to it. Agents also work on a ladder of trust: research, prepare, pay. Greg says the whole war is on the third rung, so blocking all three rungs is a mistake when the real risk lives in one.

    ▶ 5:15
  3. Agent loyalty goes to whatever worked last time, so a stuck checkout silently loses the customer

    Greg says humans are loyal to brands out of feeling or nostalgia, but an agent is loyal to “whatever worked last time.” If your site delivered the right part when promised, the agent remembers and returns. If it got stuck at checkout, it goes elsewhere and you never know it happened. Agents also never impulse buy and compare five suppliers in the time you open one tab.

    ▶ 7:21
  4. Companies block agents mainly to protect the highest-margin parts of the path to a sale, and fear of being disintermediated

    Greg grants real reasons: fraud, unclear authorization, and scrapers posing as agents. But he follows the money. Amazon earns tens of billions from ads, and an agent skips sponsored results. Airlines profit from upsells like seats, bags and card offers, and an agent books the cheapest nonstop. The deeper fear is that Meta or OpenAI would sit between store and customer. The same companies are building their own agents, like Amazon's Rufus.

    ▶ 8:52

Give your email to unlock the rest of the key points — for every episode.

Do more with Charlie

Free — sign in with your email, no password.