Tom Verrilli's retail-market framing: e-commerce has never exceeded roughly 20% of US retail spend in 30 years, despite repeated waves of hype that it would eventually dominate.
His read on agentic commerce: it fits high-intent, programmatic purchases (you already know what you want; an agent executes the transaction). Low-intent, social/discovery-driven shopping — the mode live commerce serves — is a genuinely different behavior, not a weaker version of the same thing, and the two are expected to coexist rather than one displacing the other.
Tom Verrilli (Whatnot CPO) draws the agentic/social line in terms of shopper intent rather than product category. High-intent shopping — specific, programmatic, "I know what I want" — is the part AI agents can automate well. Low-intent shopping — open-ended browsing, discovery, "show me something interesting" — depends on human judgment and social context, which is why it resists agent automation and is where live/social commerce lives.
He extends this into a claim about structural novelty: live commerce may not just be a livestream veneer on catalog e-commerce, but a genuinely new category, because it combines internet-scale reach with the same non-transactional value physical retail has always provided — a curated, social browsing experience — rather than simply digitizing a transaction that used to happen in a store.
Related: CPM Economics vs. Commerce Economics, Data Network Effect: Interest Graph vs. Friend Graph.