> ## Documentation Index
> Fetch the complete documentation index at: https://docs.navigara.com/llms.txt
> Use this file to discover all available pages before exploring further.

# AI Levels

> How Navigara places each person on one of five AI levels, and how to find the next step.

<video controls className="w-full aspect-video rounded-xl" poster="https://storage.googleapis.com/navigara-public-artifacts/docs/videos/ai-levels.jpg" src="https://storage.googleapis.com/navigara-public-artifacts/docs/videos/ai-levels.mp4">
  <track kind="captions" srcLang="en" label="English" src="/videos/ai-levels.vtt" default />
</video>

AI Levels show how a person's AI habits grow. AI can answer a quick question, or it can finish a whole task while you do something else. Navigara measures five habits from your AI tool telemetry and combines them into one of five levels.

## The five habits

Each habit has a target, and your organization sets it. A habit at or above its target scores 100%. Below the target, the score falls in proportion.

| Habit | What it measures |
| - | - |
| **Daily use** | The days a person uses any AI tool, in each week they use one. A week with no use does not count. |
| **Agent time** | Agent hours on a day the person uses AI, averaged over those days. |
| **Time before you step in** | Agent minutes for each prompt. Short prompts with constant corrections are pairing; a long run is delegation. |
| **Agents at once** | Agent minutes in a day divided by the time at least one agent ran. One agent at a time scores 0%. |
| **Work while you are away** | Agent hours on an active day that ran with no one steering. Each agent turn counts after its first 10 minutes. |

## The five levels

The five habit scores average into one adoption score. The score places the person on a level. Each level covers an equal fifth of the score range.

export const levels = [["Curious", "Tries AI now and then."], ["Regular", "Uses AI most weeks, mostly for quick questions."], ["Fluent", "AI is a daily habit."], ["Delegator", "Hands whole tasks to agents and lets them finish."], ["Orchestrator", "Runs several agents at once, and work goes on while away."]];

<div className="not-prose my-6 flex flex-col gap-2">
  {levels.map(([name, text], i) => (
      <div key={name} className="flex items-center gap-4 rounded-xl border border-gray-200 dark:border-white/10 px-4 py-3">
        <div className="flex shrink-0 items-end gap-1" aria-hidden="true">
          {[0, 1, 2, 3, 4].map((k) => (
            <span key={k} className={k <= i ? "block w-2 rounded-sm" : "block w-2 rounded-sm bg-gray-200 dark:bg-white/10"} style={{ height: `${8 + k * 4}px`, ...(k <= i ? { backgroundColor: "#6366F1" } : {}) }} />
          ))}
        </div>
        <div className="min-w-0">
          <div className="font-semibold text-gray-900 dark:text-gray-100">{i + 1}. {name}</div>
          <div className="text-sm text-gray-600 dark:text-gray-400">{text}</div>
        </div>
      </div>
    ))}
</div>

Navigara calculates the score each day over the trailing 90 days. A person with no AI activity in the window has no score and no level.

A team's level is the level of its active people's average score. A person counts in only one team: the team where they had the most active days in the window.

## Your next step

You climb one level at a time. Navigara shows your **next step**: the habit with the lowest score. A higher score on that habit moves you up most.

| Lowest habit | Next step |
| - | - |
| Daily use | Use an AI tool on more days. |
| Agent time | Hand more of your daily work to an agent. |
| Time before you step in | Give the agent a whole task in one prompt, and let it finish. |
| Agents at once | Start a second agent while the first one works. |
| Work while you are away | Start a task and move on, instead of watching the agent work. |

Work on one habit at a time, and watch it grow week by week.

## How to use levels

* **Coach, do not rank.** Levels show habits, not output or quality. Use them to find the next habit to coach.
* **Most teams do great work at Fluent and Delegator.** Orchestrator is not a goal for every role.
* **Read levels next to delivered work.** The [AI Adoption](https://app.navigara.com/adoption) page plots each person's level against their [ETV](/concepts/performance#the-etv-unit), so you can see whether stronger habits come with more delivered work.

<Note>
  Levels need AI tool telemetry. See [AI tools](/integrations/ai-tools) to connect your tools.
</Note>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.