Checking AI Readiness
On this page
Overview
AI Readiness is the Skill Assessment screen that shows how ready your organization is to work with AI. It aggregates the AI tool usage of the members being measured to show which Skills and MCP servers are used, in which repositories, and by how many members. It also shows your organization's AI Readiness score and how much of an agent's work goes into exploration.
Members with Admin Authorization or Manager Authorization can open AI Readiness.
About the beta Skill Assessment is currently offered as a beta. The items on this screen and the way they are aggregated may change.
Opening AI Readiness
- In the sidebar, click AI Readiness under Skill Assessment.
- Choose the period to aggregate: Last 7 days, Last 30 days, or Last 90 days. Last 30 days is selected when you first open the screen.
From top to bottom, the screen shows Score / breakdown, Agent task efficiency, and Readiness, followed by the Skills, MCP, and repository lists.
Score / breakdown
| Item | Description |
|---|---|
| AI Readiness score | A score out of 100: the average Environment Engineering score of the evaluated members, from their AI Collaboration Skills |
| Rank and Top ◯% | The rank of your organization's AI Readiness score and its relative position |
| ◯ evaluated | The number of evaluated members the score is based on |
| Breakdown | The score of each item that makes up Environment Engineering (Context Scaffolding, Guardrail Setup, and Tool Provisioning), and the number of members evaluated on that item |
The score is not a total of usage in the selected period; it is calculated from each member's latest evaluation. No comparison with the previous period is shown for the score.
If no member has been evaluated yet, "No member has been evaluated yet" is shown. If the breakdown is not available yet, "The breakdown has not been generated yet. It appears the next time members' reports are generated." is shown.
For the Environment Engineering axis, see What is Skill Assessment?.
Agent task efficiency
The Agent work breakdown divides the tool calls agents made in the selected period into Explore, Execute, Verify, and Other, and shows the count and share of each across repositories.
- The share is the ratio of tool call counts. Calls that could not be classified into any phase are left out.
- Previous shows the share in the preceding period of the same length. The change against the previous period is the difference in share, in points.
- Only Explore is colored to show whether the change is good or bad. The more context is in place (CLAUDE.md, external sources), the less of its time an agent spends exploring.
If no tool calls were made in the period, "No tool calls were measured in this period." is shown.
Readiness
The Skills, MCP, and Repositories cards show the number detected in the selected period and the top three entries. If the number changed from the previous period, the difference is also shown.
| Card | Count | Order of the top three |
|---|---|---|
| Skills | Number of skills that ran | Most runs first |
| MCP | Number of MCP servers that were called | Most calls first |
| Repositories | Number of repositories with usage | Best explore efficiency first |
Click a card to switch the list below to the corresponding tab.
Skills, MCP, and repository lists
Use the Skills, MCP, and Repositories tabs to switch lists. Only what was actually used during the selected period is listed. A skill or MCP server that is installed but never ran, and a repository with no work in it, do not appear.
Click a column header to sort (except the Skills and MCP columns on the Repositories tab). You can show 10, 20, or 50 rows per page (10 by default).
Skills tab
| Column | Description |
|---|---|
| Skills | The skill name, with its AI tool and source |
| Execution Count | Number of times it ran in the period |
| Users | Members who ran it / members measured in the period |
| Repositories | Repositories it ran in / repositories with usage in the period |
MCP tab
| Column | Description |
|---|---|
| MCP server | The MCP server name, with its AI tool |
| Calls | Number of times it was called in the period |
| Users | Members who called it / members measured in the period |
| Repositories | Repositories it was called from / repositories with usage in the period |
A skill or MCP server with the same name appears in separate rows when it is used from different AI tools.
Repositories tab
| Column | Description |
|---|---|
| Repositories | The repository name (in owner/repo format) |
| Explore efficiency | The share of tool calls that were not exploration. Shown as "—" when no calls could be classified |
| Skills | Skills that ran in this repository / skills that ran in the period |
| MCP | MCP servers called from this repository / MCP servers called in the period |
Rows shown in the warning color
Values and bars are shown in the warning color (orange) in the following cases. Use them as a guide to find Skills, MCP servers, and repositories with room to expand.
- Users is 25% or less of the members measured
- Repositories on the Skills and MCP tabs, or Skills and MCP on the Repositories tab, is less than 50% of the total
- Explore efficiency is 60% or less
Detail screens
Click a row in a list to open the detail screen for that skill, MCP server, or repository. You can choose the period on the detail screen as well. When you open it from a list, the period selected in the list carries over. Click Back to AI Readiness to return to the list.
Skill and MCP server details
The header shows Users, Repositories used in, and Execution Count (Calls for an MCP server). Calls to an MCP server are counted whether or not the call succeeded.
Below it are two tables.
| Table | Columns |
|---|---|
| Usage by repository | Repositories / Execution Count (Calls) / Users (compared with the number of members who worked in that repository) / Last used |
| Usage by member | Members / Execution Count (Calls) / Last used |
Usage by member lists only members who used it in the period. Click a member's row to open that member's detail screen.
Repository details
The header shows the number of Developers, Skills, and MCP, followed by the Agent work breakdown for that repository. Repository details do not show a comparison with the previous period.
Below that, the Skills and MCP tables list each skill that ran and each MCP server that was called in that repository during the period, with its Execution Count (Calls) and Users (compared with the number of members who worked in that repository).
When data appears
AI Readiness is aggregated from the AI tool usage of the members being measured. To start measuring, see What is Skill Assessment?.
- Skills, MCP, and Agent task efficiency: Shown once measured members run a skill or call an MCP server with their AI tools in the selected period.
- Figures by repository: Only usage whose repository could be identified is counted. As a result, the figures on the Repositories tab and in "Usage by repository" may be lower than the total of all usage.
- Members who use Codex: The member needs to give the approval that Codex asks for the first time it starts. Until they do, their Codex usage is not included in the figures by repository.
- AI Readiness score: Shown once members' evaluations are generated. The breakdown appears from the next time members' reports are generated.