Skip to content

AI System Dashboard ​

The AI System dashboard shows where one AI System is in its lifecycle, and how much of the work for that lifecycle stage is done. It shows the Controls that are necessary at that lifecycle stage, how many of them are Executed, and how the risk value compares with the risk limit.

What this is ​

Use the dashboard to answer these questions quickly:

  • Which Controls are necessary for this AI System at its current lifecycle stage, and how many of them are Executed?
  • What stops the work at this time?
  • Where is the risk, and how does it compare with the limits?
  • How many Controls has the Control Assessment Agent assessed, and what is their average readiness?

Where in Modulos ​

  • AI System → Dashboard

The Dashboard shows the status of the work. The registration record is on AI System → Overview: the intended purpose, the EU AI Act classification, an Owner, the Current State, and the scopes. For a Pending, Approved, or Rejected AI System, the Overview is the first item in the sidebar, and it opens from the AI Registry. A Draft opens in the registration wizard and has no Dashboard. The organization AI System has no Overview and opens on its Dashboard. See Register an AI System.

Who can do what ​

Permissions

  • All members of the AI System can see its dashboard. When governance is on for your organization, all members can also use Export in the More actions menu.
  • The Add New items of the More actions menu show only for users with the related right. For example, Framework needs the right to add frameworks, and User needs the right to give AI System roles.
  • Change Current Stage, Set Lifecycle Stage, and AI System Settings show only for users who can change the AI System: AI System Owners and the users with full access to all AI Systems. See AI System Settings.
AI System dashboard with the AI System Lifecycle card, its lifecycle stage buttons, the percentage of Executed Controls, the Controls by framework, and the summary card with the Risk Value and the AI Assessment.

The AI System dashboard shows the Controls of each lifecycle stage, the percentage that is Executed up to the current lifecycle stage, the risk value, and the AI assessment.

  1. AI System Lifecycle The Controls of each lifecycle stage, and how many of them are Executed.
  2. Current stage The lifecycle stage of the AI System. Select a different lifecycle stage to see its Controls.
  3. Controls by framework The Executed Controls and the total for each framework. Select a framework to open its Controls.
  4. Risk Value The total risk value against the risk limit of the AI System.
  5. AI Assessment The Controls that the Control Assessment Agent assessed, and their average readiness.
  6. More actions Opens the menu with Export and Add New.

How it works ​

The dashboard collects the data of the work items of the AI System. From top to bottom, it shows the AI System Lifecycle card, the summary card, the risk charts, and the Frameworks section.

More actions ​

More actions is the button with three dots at the top right of the dashboard. Its menu has these items:

  • Export makes one PDF of the AI System. It shows when governance is on for your organization. See Reports & Exports.
  • Add New is a group of items: Framework, Evidence, Asset, and User. Framework opens Settings → Frameworks. If a framework is not at its latest version, Modulos first asks you to update it in Settings.

AI lifecycle stage ​

The AI System Lifecycle card is at the top of the dashboard. It shows which Controls the AI System needs at its current lifecycle stage, and how many of them are Executed.

Not all Controls of an AI System are necessary at the same time. For example, a Control for the retirement of the system is not necessary while the system is in development. The card thus counts the Controls up to the current lifecycle stage. It shows the Controls of the future lifecycle stages, but it does not count them.

The card shows when governance is on for your organization. The organization AI System does not show it.

Lifecycle stages and the Controls that count ​

The lifecycle stages are the values of the AI System Lifecycle tag. They have this sequence:

  1. Inception
  2. Design and Development
  3. Verification and Validation
  4. Deployment
  5. Operation and Monitoring
  6. Re-Evaluate
  7. Retirement

When you select a lifecycle stage, the card shows its description. The AI Lifecycle Stage of the AI System is its current lifecycle stage, and the card marks it with "Current stage". A future lifecycle stage is one after the current lifecycle stage.

These rules set which Controls the card counts:

  • Only Executed and Not Executed Controls count. Out of Scope and Archived Controls do not count on the card.
  • A Control is in a lifecycle stage when its AI System Lifecycle tag has the name of that lifecycle stage. A Control with two of these tags is in the two lifecycle stages.
  • Continuous obligations are the Controls that count and have no AI System Lifecycle tag. They apply at all lifecycle stages, from the first to the last. On the Controls page, the Lifecycle Stage value No stage (continuous obligations) shows the Controls with no AI System Lifecycle tag. It also shows the Out of Scope and Archived ones (see Controls).
  • A Control of more than one framework counts one time in the totals of the card.

When no Control counts, the card shows one of these messages:

  • "No frameworks assigned yet": the AI System has no framework. Users who can add frameworks see Add Framework.
  • "No controls to count": all Controls of the frameworks are Out of Scope or Archived.

Lifecycle stage buttons ​

The card shows one button for each lifecycle stage, then a button for Continuous obligations. Each button shows:

  • a ring with the share of Executed Controls in that lifecycle stage only. When all its Controls are Executed, the ring is full and has a check mark. The ring of a future lifecycle stage stays empty.
  • the number of Executed Controls and the number of all Controls, for example "12 / 30"
  • the number of Not Executed Controls, for example "18 not executed". For a lifecycle stage before the current one, this number is red. A future lifecycle stage does not show it.
  • "No controls", when no Control of the lifecycle stage counts

The card selects the current lifecycle stage first. When no lifecycle stage is set, it selects all Controls. Click a lifecycle stage to select it. To go back, click it again or click Back to the current stage. When no lifecycle stage is set, click Back to all controls.

Percentage and bar ​

The large percentage shows how many of the Controls in scope are Executed. The selection sets the Controls in scope:

  • The current lifecycle stage or one before it: the Controls of that lifecycle stage and all lifecycle stages before it, plus the continuous obligations. A Control of two lifecycle stages counts one time.
  • A future lifecycle stage: the percentage stays at the current lifecycle stage. The card shows "Not counted yet. Its n controls count once the AI System reaches this stage."
  • Continuous obligations: only the continuous obligations.
  • No lifecycle stage set: all Controls that count, for each selection. The card shows "No lifecycle stage is set." with Set Lifecycle Stage.

The text next to the percentage tells what it counts, for example "controls executed up to the Current Stage".

For example, the AI System in the screenshot at the top of this page is in Operation and Monitoring, the fifth lifecycle stage. Its lifecycle stage buttons show these numbers, and the percentage counts the Controls up to the current lifecycle stage, plus the continuous obligations:

Lifecycle stageExecuted / counted ControlsIn the percentage
Inception25 / 25Yes
Design and Development44 / 47Yes
Verification and Validation12 / 12Yes
Deployment17 / 19Yes
Operation and Monitoring13 / 22Yes ("Current stage")
Re-Evaluate0 / 3No
Retirement0 / 1No
Continuous obligations2 / 3Yes
Percentage113 / 12888%

Of the 15 Not Executed Controls in scope, 9 are in the current lifecycle stage, 5 are in earlier lifecycle stages, and 1 is a continuous obligation. The Controls of Re-Evaluate and Retirement count from the time when the AI System gets to those lifecycle stages. A Control of a future lifecycle stage that is also in the current or an earlier lifecycle stage counts with that lifecycle stage. In this example, 1 of the 3 Re-Evaluate Controls is also in Operation and Monitoring, and it counts there.

When you select a future lifecycle stage, the card shows its Controls and the text "Not counted yet":

AI System Lifecycle card with the future lifecycle stage Re-Evaluate selected. The card shows Not counted yet, the View Controls button, and Back to the current stage.

With a future lifecycle stage selected, the card shows its Controls as not counted yet. The percentage stays at the current lifecycle stage.

  1. Re-Evaluate A lifecycle stage after the current one. Its ring stays empty.
  2. Not counted yet The Controls of this lifecycle stage count from the time when the AI System gets to it.
  3. View Controls Opens the Controls of this lifecycle stage.
  4. Back to the current stage Selects the current lifecycle stage again.

Below the percentage, a bar and a legend divide the Controls in scope into these groups:

  • "Executed"
  • "Not executed in the Current Stage". When you select a lifecycle stage before the current one, the row has its name, for example "Not executed in Deployment".
  • "Not executed in earlier stages". The first lifecycle stage does not have this row.
  • "Not executed continuous obligations"

With no lifecycle stage set, or for the continuous obligations, the bar and the legend show only "Executed" and "Not Executed".

Reopened by failed tests shows in red at the end of the legend. It counts the Not Executed Controls in scope that are linked to an Enabled test with the latest result Failed. It shows only when this number is more than 0. Click it to open Runtime Inspection. For how a failed test sets a Control back to Not Executed, see Results & Remediation.

Controls by framework ​

The right column of the card shows the Controls of the selection by framework. Its title names the selection, for example Controls by framework in Deployment, All controls by framework, or Continuous obligations by framework.

  • The column shows the three frameworks with the most Controls, each with its Executed Controls and its total. One more row adds up the other frameworks, for example "2 more frameworks".
  • A Control of more than one framework counts in each framework row. The card then shows "A control in several frameworks counts in each."
  • With a lifecycle stage or all Controls selected, click a framework to open its Controls on the Controls page. With a lifecycle stage selected, the Controls page also filters by that lifecycle stage.

View the Controls ​

A button below the description of the selection opens the Controls page:

  • View Not Executed Controls opens the Not Executed Controls of the selected lifecycle stage. With all Controls selected, it opens all Not Executed Controls. With the continuous obligations selected, it opens the Not Executed continuous obligations. It does not show when all Controls of the selection are Executed.
  • View Controls shows for a future lifecycle stage. It opens the Controls of that lifecycle stage.

No button shows when the selection has no Controls.

Set the lifecycle stage ​

The current lifecycle stage is the AI Lifecycle Stage field in Settings → General Settings. After registration, you change it only there. The card has a shortcut to the field: Change Current Stage when a lifecycle stage is set, and Set Lifecycle Stage when none is set. These buttons show only for users who can change the AI System (see Who can do what).

  1. On the AI System Lifecycle card, click Change Current Stage or Set Lifecycle Stage.
  2. In Settings → General Settings, select the lifecycle stage in AI Lifecycle Stage.
  3. Click Save Changes.

Summary card ​

The summary card is below the AI System Lifecycle card. It has these columns:

  • Description: the description of the AI System, or "No description yet.", with Owner(s), Registration Status, and Last updated. AI System Settings opens the settings of the AI System.
  • Risk Value: shows when risk management is on for your organization. It shows the total risk value against the risk limit of the AI System. It also shows the percentage of the limit that it uses (100% at most) and a bar. View Risks opens AI System → Risks. Until the AI System has a risk value, the column shows "No risk value yet."
  • AI Assessment: see AI Assessment.

A Rejected AI System has no risk limit. When it has a risk value, the Risk Value column shows the value and No risk limit. See Registration approval.

The organization AI System has no Registration Status and no Risk Value column.

AI Assessment ​

The AI Assessment column of the summary card shows the coverage of the Control Assessment Agent. It shows when governance and the Control Assessment Agent are on for your organization.

  • "n of all m controls assessed": the number of Controls with a stored assessment, out of all Controls that are not Archived and not Out of Scope. Executed Controls count.
  • "x% average readiness": the average readiness score of the assessed Controls. Only the latest assessment of each Control counts.
  • The bar shows the share of assessed Controls.
  • View Assessed Controls opens AI System → Controls with the assessed Controls, the lowest readiness score first.

Before the first assessment, the column shows "No controls assessed yet". For users who can run the agent, Create Job opens the Controls page with the dialog to create an assessment Job. To find the Controls with no assessment, use Filter → AI Assessment → Not assessed on the Controls page. When the AI System has a lifecycle stage, the Controls page opens at it: remove the Lifecycle Stage filter to include all lifecycle stages.

Risk signals ​

When risk management is on for your organization, the dashboard shows these sections below the summary card. The organization AI System does not have them.

  • Risks Overview: the top quantified risks by value, with a filter for one risk category, and Risk Status by Category
  • Risk Value Over Time: the risk value of the AI System in a time range, compared with its risk limit

The total risk value is in the Risk Value column of the summary card. It adds up the values of the risk threats of all risks of the AI System. See How risk becomes a portfolio signal. The risk values help you make decisions. The audit trail is on the items and in their Comments and Logs.

Frameworks ​

When governance is on for your organization, the Frameworks section shows at the bottom of the page. It has one card for each framework of the AI System. Each card shows the completed and total Requirements and Controls of the framework. See details opens the Requirements of that framework. Add Framework opens the Add Framework(s) dialog of Settings → Frameworks. If a framework of the AI System is outdated, Modulos first tells you to update it in Settings.

How to use it ​

  1. Start with the scope Make sure that the AI System description agrees with what you govern.
  2. Set the lifecycle stage Keep the current lifecycle stage correct, because it sets which Controls the dashboard counts.
  3. Find what stops the work On the lifecycle card, look for Not executed Controls up to the current lifecycle stage. Look also in the continuous obligations.

Keep to these rules:

  • Do the largest gaps first: work on the Controls and Requirements with the most work that remains.
  • Use risk to make decisions: quantify and monitor the most important threats and compare them with the limits.
  • Open the audit trail: open the items to see the rationale, the Evidence, and the logs.

Important considerations ​

  • The dashboard is a summary. Make sure that each decision is traceable through Evidence, reviews, and the logs of the items.
  • If the dashboard is not complete, the cause can be a missing scope, missing frameworks, or missing permissions.
  • The AI lifecycle stage sets which Controls the lifecycle card counts. When the AI System goes into a new lifecycle stage, set it in the settings. Until you do, the card does not count the Controls of that lifecycle stage.