In Cenedril you anchor the leadership of the AI management system in the Leadership chapter of the AI Compliance Assistant: you assign the AI roles to people, send the top-management commitment declaration for electronic signature, generate the AI policy as the guiding document and create measurable AI objectives from templates. At the end the roles are published, the charter is filed as a record, the AI policy awaits its approval under Documentation, and the objectives are ready for later measurement.
Open the chapter
Open Leadership
In the sidebar open Wizards, choose the Artificial Intelligence card and click Leadership in the chapter list. The entry page Leadership and Commitment explains why leadership commitment matters and shows the chapter’s four working pages as cards. At the top sits the progress bar with the steps Leadership & Commitment, Commitment Declaration, Roles & Responsibilities, AI Policy and AI Objectives; every step can be clicked directly.
Settle the order
Next at the bottom of the page leads to the commitment declaration. Its suggestion box takes the names from the AI roles; this guide therefore assigns the roles first and returns to the declaration afterwards. Click Roles & Responsibilities in the progress bar.
Assign the AI roles
Read the table
The AI Roles & Responsibilities page shows a table with the columns Responsibility, Accountable and Responsible. Each row is an obligation from the AI management system and the AI Act; the line below it names the source. The accountable person answers for the result, the responsible person does the work. The expander Show the authorities of the … role shows which authorities the role catalogue provides for the suggested role. The rows Data quality and AI development appear only if your organisation stated in the organisational context that it develops its own AI systems or operates its own model.
Name the AI officer
Start with the highlighted AI Officer row. It is the only row that Save & Continue checks: without an accountable person the page cannot be completed. The person named here appears by name in the generated policies, for instance in the AI policy and the AI usage policy. The AI Policy row below it holds the owner of the AI policy; that person later receives the reminder for the annual review.
Fill the remaining rows
Pick the people per row via Select a person. The selection draws on the people in your asset register; people with a linked user account are listed first. A newly typed name immediately creates a new person in the asset register. Check the authority per row: the Resources for the AI management system row only holds if the person can release budget and staff, and the Human Oversight row needs someone who may discard outputs or halt operation.
Publish the roles
Save as Draft stores any interim state without a check. Save & Continue checks the AI officer row, sends every person with a user account who is newly entered in a row a notification and an email with their responsibilities, and moves on to the AI policy. For the commitment declaration, click Commitment Declaration in the progress bar.
Send the commitment declaration for signature
Check the preview
The Leadership Commitment Declaration page shows the text of the declaration at the top, together with the Scope of the AI management system and the Requirements of interested parties from the Context & Scope chapter. Both travel into the later charter record; they can only be changed where they were recorded. The link View AI scope statement appears once the scope statement exists as a document in the documentation area.
Add signatories
The yellow box Recommended signatories from the AI roles lists the people entered as accountable or responsible in the roles who are not yet on the list; one click on a name takes it into the Signatories list. Once every suggestion has been adopted, the box disappears. Add further people under Add person with Name, Email and Role and confirm with Add. The intended signatories are top management: the managing directors or the board and the leaders who carry supervisory responsibility for the AI management system. The AI officer signs along if they belong to leadership.
Complete the row and send
Adopted suggestions bring neither role nor email address. Fill in Role and Email in the row and tick Top management for the members of leadership. As soon as an email address is entered, Send for signature becomes active; the click sends the invitation immediately, saving the page is not required for that. The person opens the link from the email, requests a confirmation code and signs in the browser. This produces an advanced electronic signature under the eIDAS Regulation. While an invitation is running, role and address in the row are fixed.
Track the status
Each row shows the state of the invitation as a badge: Sent, Viewed (the link was opened), Signed or Expired. As long as nothing is signed, the icons on the right offer Resend and Revoke; after a revocation the row returns to its initial state with the fields and the send button. After the signature, View record leads to that person’s individual record. The counter above the list shows how many people have confirmed.
Create the charter
Save as Draft stores the list of people without a check. Save & Continue requires at least one entered person, creates the aggregated AI commitment charter as a record and moves on to the AI roles. Once the charter exists, the green notice with View charter appears at the top; it is filed under Documentation → Records. Every later signature updates the charter. Then switch to AI Policy via the progress bar.
Generate the AI policy
Generate the draft
The AI Policy page shows the document editor with the language flags under Select the language of the document:. Choose the language and click Generate Document. The draft is built from the organisational context, the AI roles, the AI register and the risk scale; if a source is missing, the generator inserts a general fallback sentence. Once the displayed language has content, the button reads Regenerate; after a confirmation it replaces the entire text of the displayed language, your own edits included. Both buttons act only on the displayed language.
Edit the draft
Four places deserve a deliberate check: the paragraph on the role (provider and deployer, or deployer only), the list of roles with the names from the AI roles, the related documents, and the references to the information security policy and the data protection documentation. Remove commitments your organisation will foreseeably not honour. The section on the AI objectives deliberately stays general; the measurable objectives follow in the next step.
Set up the annual review
Below the document sits the card Annual review of the AI policy. Set reminder (yearly) creates a recurring date; the Recipients are prefilled from the AI Policy row of the AI roles. The substantive review belongs in the management review.
Publish the policy
Save as Draft stores both languages without publishing anything. Save & Continue publishes them, creates one version per language with content under Documentation → Policies & Procedures for review and approval, and moves on to the AI objectives. An editor that is empty in both languages cannot be published. The completion flag of the policy is set later in the Policies & Documents hub with Complete.
Create the AI objectives
Adopt templates
The AI objectives page shows the templates under Suggestions, grouped by topic: Accountability and competence, Transparency and privacy, Data quality and fairness, Robustness and security and Sustainability. If your organisation develops or operates its own AI systems, the group Objectives for responsible development additionally sits at the top. Expand a group and click Add objective in the suggestion row: a card appears at the top with title, description, What will be done? and Resources required, the template disappears from the list and the card carries the From template badge. The Add objective button above the list creates an empty card.
Complete the card
Two fields stay empty on every adoption: the Responsible person and the date. Set both right after adopting. The Type of objective toggle distinguishes One-off (achieved once, then fulfilled; field Target date) and Ongoing (held continuously and measured at a set frequency; field To be met first by). The Status describes the lifecycle: Draft, Active or Discontinued. What the objective is measured against is set later in the measurement plan on the Set up monitoring & measurement page.
Publish the objectives
Save as Draft stores any interim state. Save & Continue requires at least one objective with a title and a date and leads back to the overview of the AI Compliance Assistant, where the chapter list shows the state of the Leadership chapter.
Result: the AI roles are published and the named people notified, the invitations to the commitment declaration are sent and the charter is filed as a record under Documentation → Records, the AI policy awaits review and approval under Documentation → Policies & Procedures, and the AI objectives stand ready for the measurement plan with a responsible person and a date.