Skip to content
El Observatorio

Building & Agents · 1 min de lectura

Customer Development Is Still Your Job — Agents Can't Do It

Steve Blank's The Four Steps to the Epiphany made a foundational argument: product development and customer development are separate disciplines, and confusing them is how companies die. Discovery, validation, creation, company-building — each step earns the right to the next. That logic is more urgent now, not less, because agents make it dangerously easy to scale before you have earned anything.

Con · estudiado y reformulado para builders AI-native“The Four Steps to the Epiphany” — Steve Blank

Discovery Still Requires a Human in the Room

Blank's customer discovery phase is about getting out of the building and testing whether the problem you think exists actually exists. An agent cannot do this. It can synthesize interviews you have already run, structure notes, or draft follow-up questions — but the judgment about whether a customer's hesitation is a real signal or polite deflection is irreducibly human. Founders who delegate early discovery to automated pipelines are skipping the step, not accelerating it.

Validation Before You Build the Workflow

Blank insists you validate demand before you build the full machine. That principle maps cleanly onto agent design: build your evaluation harness before you build the agent, let real failure cases drive the architecture, and treat every workflow as a hypothesis under test. A workflow you ship before validating its outputs is exactly the kind of premature scaling Blank warns against — just expressed in infrastructure rather than headcount.

Scaling Is a Reward, Not a Starting Point

The four-step sequence is a discipline against premature scaling, and AI-native companies face a specific version of that temptation: because agents are cheap to deploy, founders reach for growth leverage before the customer and the workflow are actually proven. Blank's company-building phase — where you finally pour resources in — only arrives after the earlier steps close. The same gate applies here. Earn the repeatable customer motion first; then design the agents to scale it.

Taste and Judgment Remain the Founder's Work

What Blank identified as the founder's irreplaceable role — synthesizing messy market signals into a coherent direction — has not been automated away. Agents follow paths; founders choose which paths matter. Customer development is ultimately an act of judgment about what is true, what is worth building, and what trust needs to be earned. The tools change the speed of iteration, but the founder still has to look at contradictory evidence and decide what it means.

Prueba el Desktop OS

Explora la escena startup de Atlanta, en vivo.

Todo lo que acabas de leer vive dentro de nuestro Desktop OS — un espacio de trabajo interactivo que se abre directamente en tu navegador. Su mesa de Atlanta mapea la escena startup de la ciudad: quién está construyendo, qué está pasando y dónde. Ábrelo y échale un vistazo.

Abre el Desktop OS →