Y Combinator’s motto, “Create something people want,” has never been about the product. It was about distribution. Build what fits the product and the market, and people come. For 20 years, this distribution took place via search engines, social algorithms, paid advertising and word of mouth. All human-mediated channels.
That has changed. Over the past six months, Cloudflare, Shopify, Stripe, Supabase, Netlify and Google have each invested in become ready to be an agentbuilding a new distribution channel: AI agents that visit websitesextract information, compare options, and make trades on behalf of the humans who sent them. No one coordinated this. Six companies from different sectors saw the same thing and built their project independently.
When six companies from different industries build independently for the same class of visitors, the channel is real.
Your website should always create something that people want. But officers now know how many of those people are finding it. A website that works for humans and fails for agents is a product with a flawed distribution channel.
6 companies in 6 industries made the same bet on agent infrastructure
Cloudflare dedicated an entire launch week to agents in April 2026. Not a single feature announcement. A week. Agent identity via Web Bot Auth with GoDaddy. Agent-readable content via Markdown for agents. Functions callable by the agent via WebMCP in browser runtime. Agent measurement across the Agent Readiness Score at isitagentready.com. An infrastructure provider doesn’t clear its schedule for something speculative.
Shopify delivered the Agent Toolkit so that any AI agent can browse a store’s catalog, check inventory, and complete payment through a structured API. The trader is not building anything new. Google went further with Universal Trade Protocolexpanded it to I/O 2026 with Universal Cart, then added the Agent Payments Protocol to the FIDO alliance with 60 organizations. The commercial layer for agents went from draft specification to production integration in less than three months.
Stripe has delivered Projects, a platform where AI agents can create accounts, purchase domains, deploy infrastructure, and manage subscriptions through Stripe’s payment rails. The infrastructure purchase layer became negotiable by agents overnight.
Netlify builds netlify.aia dedicated surface where AI agents can deploy websites, manage projects and access the full platform using agent skills. This is not a feature built into the existing product. A separate entry point designed for non-human visitors. I spoke to their CEO about it last weekand the reasoning was simple: if agents are going to deploy websites on your platform, give them a front door designed for them.
And then there is Supabase. Their tagline is “Postgres Development Platform”. This means almost nothing to the mood coders who have made Supabase the default database for AI-created apps. But it’s a perfect, machine-readable description of what Supabase is and what a coding agent can do with it. The tagline reads like it was written for an agent, not a human. Whether Supabase designed the slogan for agents is irrelevant. The machine-readable identity worked.
None of these companies responded to each other. They all addressed the same thing: a class of visitors that needs a machine-readable identity, structured content, discoverable actions, and predictable transaction flows. A class of visitors that grows because the humans behind the agents prefer it.
What it really means to invest in agent preparation
Agent preparation is not a budget line or a new team to hire. Agent readiness is a set of infrastructure decisions about how your website delivers what it does to non-human visitors.
Can agents read your content? If your website depends on JavaScript rendering To view their main information, most agents see a blank page. Server-rendered HTML with semantic structure is the basis. Server-rendered HTML is the easiest gap to fill and the one with the most immediate impact.
Can agents find out what you offer? A robots.txt which recognizes AI user agentsa site map that remains up to date, structured data that names your entities and relationships. These are fundamentals of the classic web applied to a different class of visitors. Nothing here is new. The visitor is.
Can agents act? If your website sells something, can an agent complete the purchase? If your website provides a service, can an agent invoke it? The protocol layer (UCP, PCMWebMCP) is where it is. Most websites aren’t there yet. It’s good. But the companies here are building the gap.
The Moat takes it seriously before your competitors. Shopify already enables access to Agent Toolkit by default for all merchants. Those who pay attention to what this means for their product data and checkout flow are the ones capturing agent-referred traffic. Early SaaS products with WebMCP tools enable agent-discovered users. Supabase is the default database for agent-created applications because its identity was machine-readable before anyone thought of it. The window is open because most websites have not started.
“Create something people want” now includes agents
YC’s motto has not changed. But agents are now involved in how people find what you’ve made, compare it to alternatives, and decide whether or not to buy it. A website that only works for humans is a product with a distribution problem; he doesn’t know it yet.
More resources:
This article was originally published on No hacks.
Featured image: Roman Samborskyi/Shutterstock
