← All retail signals
Retail

Post-Interface Retail

You name the outcome and the machine handles the rest — no cart, no checkout, just the wanting and the arriving.

What is it?

Shopping stops being an interface and becomes an instruction: you name the outcome, an agent sources, compares and buys it, and the storefront turns into something machines parse rather than something people wander through. Whether that instruction still needs trusting is the open part.

Why now?

Retailers spent decades building storefronts to hold attention long enough to convert it; an agent that already knows what a shopper wants skips that funnel entirely, going straight from stated intent to purchase. Instant Checkout made the plumbing real, not hypothetical. The force underneath is disintermediation, not convenience.

Who’s building it?

Model providers and payment rails are building the pipes; merchants are still deciding whether to plug in.

Assistant makersPayment rail providersMerchant platformsEarly-adopter retailersRegulators · FTC
The full read
Why now, who is building it, where the market goes, and what stays unsettled.
Get weekly →
01 · The market
Who’s in it
Builders — assistant makers and payment rails, merchants plugging in.
Buyers — shoppers trading browsing for stated intent, starting with routine, low-consideration purchases.
Gatekeepers — platforms and merchants deciding whether to open their catalogue to agents.
What’s at stake
The margin on being found at all — discovery moves from a shopper’s eye to an agent’s ranking logic.
Which retailers stay visible once shoppers stop browsing.
Whether checkout data stays merchant-owned, or passes to the platform.
What’s unsettled
Whether outcome-shopping stays for errands or spreads to considered purchases.
Who ranks a merchant first — the platform or the agent itself.
Whether merchants can profitably serve an audience that no longer visits their site.
02 · How the demand matureda progression, not a timeline
Novelty
Assistants handle simple, replaceable purchases first — low-risk, easy-to-reverse trial runs.
Habit
Repeat orders and errands move to agents once one purchase goes cleanly.
Trust
Bigger, considered purchases wait on trust — verified sourcing, error and refund handling proven out.
Category
Enough volume routes through agents that merchants build for the agent, not the eye.
03 · Where the market goesthe forks, not a prediction
→Open protocol or platform lock-in — whether agentic checkout stays an interoperable standard, or one company’s ecosystem absorbs it and decides who else can build.
→Merchants may optimise listings for machine parsing well before shoppers change their habits.
→Discovery could concentrate in one dominant assistant, or fragment across many rival agents instead.
→Payment and buyer data may settle with merchants, or with the platforms between them.
The open question is whether shoppers actually want to hand over the choosing, or only the doing — convenience is proven, delegated taste is not.
04 · The Agentic Delegation deskadjacent markets, moving now
Agentic Commerce
The open protocol behind agent checkout.
OpenAI and Stripe launch open Agentic Commerce Protocol.
CNBC
The Prompt Native
Shoppers now ask the assistant, not the store.
ChatGPT processes about 50 million shopping queries daily.
Digiday
The Attention Auction
Forecasts put a fifth of retail as agentic.
Agentic commerce could reach 15-25% of US retail.
Bain & Company
05 · The Wire
01
ChatGPT launched Instant Checkout with Etsy, Shopify (Sept 2025).
Marketing Dive ↗
02
AI agents like Meta’s Muse order groceries and book for users
TechCrunch ↗
03
ChatGPT shifts to product discovery; merchants use their own checkout
CNBC ↗
The panel on this signal
They agreed on
–All four agreed the plumbing for agent-led buying is real, working, and already shipping into stores.
–Trust, far more than technology, gates any deeper adoption.
–Merchants are optimising for machines well before shoppers change their own habits.
The retail analyst disagreed on
–A single launch involving two retail partners is a limited pilot, not evidence sufficient to support claims of a category-wide reorganisation of how retail transactions are structured across the broader market.
Know more about the method →
✦How this signal was built
01 · Panel
A Delphi panel of 4 AI analysts, each reasoning independently.
→
02 · Debate
They challenge and revise each other’s reads over two rounds.
→
03 · Evidence
Every claim is tied to a cited, linked source on The Wire.
→
04 · Synthesis
A moderator writes it into one read, disagreement and all.
✦Signal changelogre-read as evidence changed
2025
Shopping collapses into stated intent: an agent sources, compares and buys, and the cart and checkout fade.
→
2026
In ChatGPT, assistants take over product discovery while merchants use their own checkout; OpenAI pivoted away from in-chat Instant Checkout. CNBC ↗