The go-to-market framework where AI agents discover, trial, and adopt software — before the human ever logs in.
Explore the FrameworkAgent-Led Growth (ALG) is a go-to-market framework developed by Enzo Duit where autonomous AI agents — not humans — are the first users of a software product. The agent discovers the product, uses its API to complete a task, and delivers the result to the human. The human sees the value before they ever create an account. ALG is the successor to Product-Led Growth (PLG) for the agentic era.
Three generations of go-to-market. Each one removed more friction. ALG removes the human from the critical discovery-to-value path entirely.
Coding agents like Cursor, Claude Code, Codex, and OpenClaw actively search for APIs and tools to solve tasks autonomously. When an agent is asked to segment a database, analyse sentiment, or send a report, it searches for the right API, calls it, and returns the result — the human only sees the final output.
This means the agent evaluates and selects tools on behalf of the human. If your product is not discoverable by an agent — through an llms.txt file, a clean OpenAPI spec, or GEO-optimized documentation — it simply doesn't exist in the agent's world. The buying decision happens before a human ever opens a browser tab.
Products that optimize for agent discovery are winning adoption faster than any ad budget or PLG motion can achieve.
Agents find you via llms.txt, OpenAPI specs, and GEO-optimized content. If you're not in the agent's context window, you don't exist. Structured, machine-readable documentation is your new SEO.
A temp-token model gives agents N free API calls with no account required. Session-hash based, not IP-based. After the trial exhausts, it returns a partial result plus a magic link for the human to upgrade.
The agent delivers the task result to the human with an upgrade URL pre-filled with context. One click, no form. The human converts because they already see the value in their hands.
PLG replaced sales-led growth by letting the product sell itself. ALG goes further — the agent completes the task before the human even evaluates.
| PLG | ALG | |
|---|---|---|
| First user | Human | AI Agent |
| Discovery | SEO, ads, word of mouth | GEO, llms.txt, OpenAPI specs |
| Trial trigger | Human clicks "Try free" | Agent calls API for a task |
| Conversion moment | Human sees product value | Human sees task result |
| Friction | Low | Near zero |
| Lock-in | Feature adoption | Agent memory + repeated use |