101 · 15 canons

Agent-led growth

How discovery, evaluation, buying and product use change when software agents become a source of demand instead of merely another acquisition channel.

TL;DR: Agent-led growth means earning selection inside agent workflows. Make the product legible and useful to agents, support verifiable transactions, and measure voluntary execution—not crawler traffic, registration or demos—as adoption.

Curated by Ali Abouelatta
Updated Sep 20, 2026

Read as Markdown

Voices in this 101

Reading path 1

See the distribution shift

These readings explain why assistants and agents could become a new layer between customers and products—and where that layer gains power.

01
AI agents will change how we use computers
Bill Gates
Microsoft cofounder and early writer on agents as the next software platform.
TL;DR
Personal agents could replace application-by-application navigation by learning preferences, coordinating services and acting across software on a person's behalf.
What you'll read
A platform thesis spanning personal life, work, health and education, plus the privacy and market questions an agent layer creates.
02
AI and the Big Five
Ben Thompson
Author of Stratechery and analyst of platforms, aggregation and advertising.
TL;DR
AI changes the strategic position of major platforms according to their control of models, clouds, distribution, devices and existing user demand.
What you'll read
A platform-by-platform analysis that helps founders ask who owns the user relationship when an agent becomes the interface.
03
Introducing ChatGPT agent
OpenAI
Builder of ChatGPT, Codex, Site Tools, workspace agents and ChatGPT ads.
TL;DR
ChatGPT agent combines research and action, moving the assistant from answering questions toward completing multi-step tasks with user oversight.
What you'll read
The consumer product surface, tools, task examples, confirmations and safety boundaries of a major agent distribution channel.
04
Bye bye, bots: say hello to AI agents
Sierra
Customer-service agent company publishing primary product and operating lessons.
TL;DR
An agent should own a goal, use systems and resolve work; a chatbot that only produces language leaves the operational burden with the customer.
What you'll read
A builder's distinction between conversational UI and company agents that can act, persist and represent policy.
05
The present and future of growth
Casey Winters
Growth operator and advisor who led growth at Pinterest and Grubhub.
TL;DR
Growth practice evolves when channels mature and product interfaces change; teams need new compounding systems rather than another round of the old playbook.
What you'll read
A growth operator's view of channel saturation, product-led growth and the organizational shift required when distribution changes.

Reading path 2

Become useful in the agent's path

Agent distribution is earned at the point of work: clear facts, callable actions, merchant data, inventory and a transaction the agent can safely complete.

06
Agentic Commerce Protocol
OpenAI
Builder of ChatGPT, Codex, Site Tools, workspace agents and ChatGPT ads.
TL;DR
ACP gives merchants and agents a shared transaction contract so discovery can become checkout without handing the entire customer relationship to one marketplace.
What you'll read
The protocol roles, merchant-of-record boundary, product and payment flow, and the open-source strategy behind agentic checkout.
07
Buy it in ChatGPT
OpenAI
Builder of ChatGPT, Codex, Site Tools, workspace agents and ChatGPT ads.
TL;DR
ChatGPT shopping turns product answers into a purchase surface, making feed quality, merchant terms and checkout compatibility part of agent distribution.
What you'll read
The customer experience, merchant integration and product-discovery surface through which agents can influence and complete a purchase.
08
Stripe's Agentic Commerce Suite
Stripe
Payments platform publishing primary work on agentic commerce and internal coding agents.
TL;DR
Agentic commerce needs more than a payment button: merchants require identity, catalog, checkout, tax, fraud and post-purchase operations that agents can call safely.
What you'll read
Stripe's primary product architecture for transactions initiated through agents and the merchant controls surrounding them.
09
Universal Commerce Protocol
Google
Builder of Gemini, Project Mariner, AP2 and major agent surfaces.
TL;DR
UCP proposes interoperable commerce building blocks so agents can discover merchant capabilities and coordinate shopping workflows without bespoke integrations.
What you'll read
Google's protocol model, capability negotiation and the relationship between agent interfaces, merchants and payment providers.
10
Agentic commerce on Shopify
Shopify
Commerce platform building merchant distribution into agentic shopping surfaces.
TL;DR
Merchant distribution increasingly includes AI shopping interfaces, but product data, availability, brand control and checkout remain operational responsibilities.
What you'll read
A commerce-platform view of how brands should prepare catalogs, storefront systems and customer relationships for agent-mediated buying.

Reading path 3

Measure demand, not agent-shaped noise

The hard part is evidence. A crawler read, a tool registration and a scripted test are not a buyer choosing the product.

11
Content Independence Day
Cloudflare
Internet infrastructure company reshaping crawler access and AI-content economics.
TL;DR
AI crawling creates a distribution and bargaining problem for publishers because content consumption can increase while referral traffic disappears.
What you'll read
Cloudflare's crawler controls, publisher argument and the distinction between access to content and reciprocal distribution.
12
Introducing Pay Per Crawl
Cloudflare
Internet infrastructure company reshaping crawler access and AI-content economics.
TL;DR
Pay Per Crawl experiments with a machine-readable price on AI access, turning an invisible fetch into an explicit economic transaction.
What you'll read
The protocol experiment, crawler identity, pricing controls and the unresolved market questions around paid content access.
13
The crawl before the fall of referrals
Cloudflare
Internet infrastructure company reshaping crawler access and AI-content economics.
TL;DR
Crawler request volume varies sharply by bot and does not reveal whether an assistant cited, recommended or transacted with a product.
What you'll read
Large-network observations of crawler behavior and the measurement boundary between HTTP access and downstream agent outcomes.
14
The crawl-to-click gap
Cloudflare
Internet infrastructure company reshaping crawler access and AI-content economics.
TL;DR
More AI crawling can coincide with fewer referrals, so publishers need separate measures for access, inclusion, citation, action and value.
What you'll read
A traffic-level argument for why traditional referral dashboards cannot fully price or evaluate agent-mediated distribution.
15
Building a growth process: first tactics second
Brian Balfour
Founder of Reforge and former VP of Growth at HubSpot.
TL;DR
A new channel should enter a disciplined loop of hypothesis, experiment, measurement and learning—not become a collection of unverified launch tactics.
What you'll read
A durable growth operating system you can apply to agent discovery without mistaking novelty for product-market evidence.

How 101 works

Lazyweb summarizes the source and tells you why to read it. The original author owns the argument. Every canon links to the original.