How AI shopping agents pick brands: Patagonia's bot-block lost the sale to REI
omooretweets · x · 2026-10-19
A marketer argues AI discoverability / agentic commerce should be every brand's top priority, and shares a free hands-on method:
- Create dummy accounts on agent platforms (Muse, Instinct, Dots) using a fresh email (optionally a VPN), ask the agent to run a search where your product should appear (e.g. "find me a winter coat for SF"), and watch its browser as it decides.
- At each stage interrogate it: where did the initial options come from? What failures/blocks eliminated choices? What variables drove the final ranking? Follow through to purchase.
In the author's Muse test: Patagonia's bot-block pushed the order to another retailer; the agent built candidates from training data (Wirecutter-style roundups, Reddit, forums), then Meta Catalog merchants, then live browser search, dropping SEO-spam and fake-review sites; REI won by having no bot challenges, verifiable inventory, same-day pickup it wasn't asked for, plus a discount.
More from Apps
- Design eng tip: use Playwright to measure loaded UI so skeletons prevent layout jank — aidenybai · 2026-10-19
- AI solves the n=1 software problem: tailored apps will replace configurable ones — sloppenheimer · 2026-10-19
- PromptShipped tracks 175 live vibe-coded apps: Claude 43, Grok 41, Lovable 41 — MicahBerkley · 2026-10-19
- Muse's side threads shine as users find OpenAI's Dot memory disorienting — manosaie · 2026-10-19
- Personal Agent Tomo Shows Off Office 'Love Wall' as It Claims Hundreds of Thousands of Users — Jackyhuang · 2026-10-19
- Pixel-art "lil' Wisps" characters revealed in Savi AI design project — AIandDesign · 2026-10-19