3 Levels to AI

I’ve been thinking a lot about how AI actually shows up in daily work. It's helpful to categorize the technology into three distinct levels of utility.

Level 1: AI Tools

This is where most people start. You prompt, it responds. Whether it's ChatGPT, Claude, or Midjourney, you are the pilot. The AI is a powerful calculator for words or images. It requires constant human input to produce anything of value.

Level 2: AI Automation

This is where you start to gain leverage. Rules + Triggers. Think Zapier with an AI step. When X happens, the AI processes Y and sends it to Z. It’s more efficient than a simple tool because it runs in the background, but it still follows a rigid, predefined path.

Level 3: AI Agents

This is the frontier. Observe, plan, act on their own. Agents don't just follow rules; they understand objectives. They can look at a problem, decide which tools to use, and execute a multi-step plan autonomously. They handle the "messy" parts of work that don't fit into a simple "if-this-then-that" logic.

The Shift in Value

Tools are useful. Automation adds leverage. But Agents are where things start to run without constant human friction. This is the most important shift in how work gets done in the modern era.

Moving Up the Stack

As a product leader or founder, your goal should be to move your workflows from Level 1 to Level 3. The more you can move towards autonomous agents, the more time you free up for high-leverage creative and strategic thinking.