Ecommerce Winners Are AI Losers
Advantages of companies that won web 2.0 leave them at greatest risk of dis-intermediation by AI super-apps.
The companies that won web 2.0 are the most exposed to AI disruption.
Looming danger
Direct insurers, online brokerages, DTC brands, neobanks, online travel - they won by executing digital channels the incumbents ignored. They were willing to operate at boundaries incumbents wouldn't touch: compliance boundaries, regulatory friction, owning the full UX. The customers who found them were tech-savvy, internet-native, self-serve. People who searched, compared, read forums, built spreadsheets.
That customer is the first to move to AI.
In ChatGPT's world, the ecommerce winner's advantages invert. Content that ranked on Google is training data. Calculators and comparison tools become tools for AI agents. Support playbooks collapse into a single AI query. The corporate muscle that built the lead on indexable, self-serve channels is the same muscle that hands the moat to a model.
Incumbents can afford to be luddites. Their customers were never going to Google anyway. Ecommerce winners don't have that insurance.
Dis-intermediation happens at both ends. AI-search gate-keeps the pre-purchase journey: users ask, AI answers, decisions are made before or without clickthrough. AI assistants gate-keep the post-purchase relationship: they read the warmup emails, retrieve policies, draft the tax return. A web 2.0 product becomes a backend.
A way out
Rebuild content for AI, not humans. Opinions, product facts, knowledge bases: restructure them to be easy for models to ingest, fact-check, and update. Presenting content in a human-centric fashion is now less important than presenting it to AI. Most users aren't clicking through anymore. That's what the AI is for.
Double down on the human-judgment layer. Ask what your best people do that a model can't yet. For a direct insurer, it's trust at the claims moment, family-planning guidance, navigating a health scare. For an online travel brand, it's the concierge who rebooks when plans change. Build products around this judgment, accountability, and relationship.
Flex the muscle that won web 2.0. Ecommerce winners got there by operating where incumbents wouldn't. The classic example: treading the compliance boundary tactfully in regulated categories. That same willingness is what builds new AI-native experiences. Be the brand an AI agent routes to by default, and the one a user picks up the phone for when the stakes are high.
Using AI to automate internal ops is table stakes. It yields cost savings and marginal UX gains. Everyone will do it.
Building new user experiences is existential.
The web 2.0 winners have a choice: figure out AI-native user experiences, or watch their channel advantages become someone else's training data.