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

# AI Proficiency

> Measure how well your employees prompt AI agents, turning prompt quality into a single trackable score.

**AI Proficiency** measures how effectively employees use AI to complete tasks.

Traditional adoption metrics, such as active users or conversation volume, tell you whether employees are using AI. AI Proficiency focuses on **how effectively they work with it**, based on the quality of their prompts and how efficiently they reach useful results.

Use AI Proficiency to understand how your organization uses AI, compare performance with external benchmarks, identify differences across teams, and track progress over time.

![Image](https://mintcdn.com/nebulyai/tyxr6Gd4zuVUslkl/images/image-53.png?fit=max\&auto=format\&n=tyxr6Gd4zuVUslkl\&q=85\&s=64361574afaea6b72d05ec1b564831d8)

## How the AI Proficiency score works

Each interaction is evaluated across **six prompting behaviors**. The behavior scores are weighted and combined into an **AI Proficiency score from 0 to 100**.

A higher score indicates that users communicate their goals more clearly, provide the information AI needs, and reach useful results more efficiently.

The AI Proficiency score is calculated over a **rolling 30-day window**. Nebuly uses all anonymized interactions from the most recent 30 days to calculate the score for your organization, team, or other selected group.

Using a rolling window keeps the score focused on **current behavior**. As new interactions are added and older ones fall outside the 30-day window, the score updates to reflect the latest snapshot of your organization's AI proficiency.

The AI Proficiency formula:

| Behavior | Weight | What it measures |
| :- | :- | :- |
| **Task clarity** | 30% | How clearly the user defines the goal of the task |
| **Context provision** | 25% | Whether the user provides the relevant background and information needed to complete the task |
| **Output specification** | 20% | Whether the user specifies the expected format, structure, or level of detail |
| **Prompt length** | 10% | Whether the prompt provides enough information while remaining focused |
| **Conversation efficiency** | 10% | How efficiently the user reaches a useful result across the conversation |
| **Constraint specification** | 5% | Whether the user provides relevant requirements, boundaries, or exclusions |

Each behavior contributes to the final score according to its weight. Behaviors with a higher weight have a larger effect on the overall AI Proficiency score.

Click **How it's calculated** to see the individual behavior scores, their weights, and how many points each behavior contributes to the final score.

## AI Proficiency score

The **AI Proficiency score** gives you an overall view of how effectively users in your organization work with AI.

The report shows the current score and how it has changed compared with the previous period.

You can also compare your organization's score with the **Nebuly benchmark**:

* **Benchmark median** shows the midpoint of the benchmark population.
* **Benchmark top quartile** shows the level reached by the top-performing quarter of the benchmark population.

These benchmarks give you context for your organization's score beyond changes in your own historical performance.

#### How the benchmark is built

The Nebuly benchmark combines **anonymized proficiency data from Nebuly customers who agree to contribute their data** with findings from **broader industry and academic research on AI proficiency and fluency**.

Using both sources gives you a broader reference point for understanding how your organization's AI proficiency compares with other organizations and with patterns observed across the wider workforce.

Benchmarks may evolve over time as Nebuly incorporates more participating organizations and new research.

## AI Proficiency over time

The **AI Proficiency over time** chart shows how your organization's score changes throughout the selected period.

It includes:

* Your organization's **AI Proficiency score**
* The **benchmark median**
* The **25th to 75th percentile range** of the benchmark

Use this view to see whether proficiency is improving, declining, or remaining stable, and how those changes compare with the benchmark.

## AI Proficiency groups

Users are automatically assigned to one of four **AI Proficiency groups** based on their proficiency:

<table>
  <colgroup>
    <col width="115" />

    <col width="726" />
  </colgroup>

  <thead>
    <tr>
      <th>Group</th>
      <th>What it means</th>
    </tr>
  </thead>

  <tbody>
    <tr>
      <td>**Expert**</td>
      <td>Consistently provides clear context, constraints, and direction to reach strong results quickly</td>
    </tr>

    <tr>
      <td>**Practitioner**</td>
      <td>Handles most AI tasks well, with occasional gaps in context or structure</td>
    </tr>

    <tr>
      <td>**Beginner**</td>
      <td>Understands the task but often leaves out useful context, detail, or guidance</td>
    </tr>

    <tr>
      <td>**Novice**</td>
      <td>Relies on simpler prompts and typically needs more back-and-forth to reach a useful result</td>
    </tr>
  </tbody>
</table>

For each group, you can see:

* The **share of users** in the group
* The **number of users** in the group
* How those users are distributed across **teams**

This gives you a quick view of the proficiency mix across your organization and where different teams sit within it.

## AI Proficiency groups progress over time

The **AI Proficiency groups progress over time** section shows how users move between proficiency groups.

Group movement is calculated over a **rolling 30-day window**. A user's current proficiency group is compared with their group 30 days earlier to determine whether they moved up, slipped back, or stayed at the same level.

Users are divided into three categories:

* **Moved up**: users who progressed to a higher proficiency group
* **Slipped back**: users who moved to a lower proficiency group
* **Stayed**: users who remained in the same group

The chart also shows how the size of each proficiency group changes over time.

Use this view to understand whether changes in the overall AI Proficiency score are translating into users progressing to higher proficiency groups.

## Build your own reports

AI Proficiency metrics and dimensions are also available in Nebuly's reporting engine.

You can combine **AI Proficiency scores**, **proficiency groups**, and individual **prompting behaviors** with your existing filters and groupings to analyze proficiency across teams, projects, business units, or other user segments.

For example, you can use custom reports to understand:

* How proficiency changes over time
* How proficiency differs across teams or business units
* Which prompting behaviors have the largest gaps
* Which groups are improving after training or enablement initiatives


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