Technical Documentation — Art. 11 & Annex IV Compliance
Article 11 of the EU AI Act requires providers of high-risk AI systems to draw up technical documentation before the system is placed on the market or put into service, and to keep it up to date throughout the system's lifecycle. Annex IV specifies the detailed content requirements for this documentation. The Technical Documentation module in Venvera provides a structured register for creating, managing, versioning, and tracking completion of all required documentation. This article covers the list view, creation form, detail page, editing, and completion tracking in full detail.
List View
Step 1 — Open the Technical Documentation List
Navigate to EU AI Act → Technical Documentation from the sidebar. The list view shows all technical documents across all AI systems in your tenant, sorted by last-updated date. Each row displays the document title, linked AI system name, document type, status badge, completion percentage bar, and version number.
Step 2 — Search for Documents
Use the search bar to filter documents by title or content keywords. The search performs a case-insensitive partial match across the document title and the linked AI system name. For example, searching "risk" will match documents titled "Risk Management Process" as well as documents linked to "Credit Risk Scoring Model".
Step 3 — Filter by Status
Use the Status filter dropdown to narrow results. Available statuses include:
- Draft — Document is being written or is incomplete. Draft documents are visible in the list but are not considered ready for audit or conformity assessment purposes.
- In Review — Document has been submitted for internal review. A designated reviewer should validate the content for accuracy and completeness before approval.
- Approved — Document has been reviewed and formally approved. Approved documents are considered audit-ready and can be presented to supervisory authorities upon request.
- Archived — Document is no longer current but is retained for historical reference. Archived documents are typically previous versions that have been superseded by newer versions.
Step 4 — Filter by Document Type
Use the Type filter to show only documents of a specific type. This is useful when you want to verify completeness — for example, filtering by "Risk Management" to see if every high-risk AI system has a corresponding risk management document.
Document Types
The module supports nine document types, aligned with the content categories specified in Annex IV of the EU AI Act:
| Type | Annex IV Section | Description |
|---|---|---|
| System Description | §1 | A general description of the AI system including its intended purpose, the name and contact details of the provider, how the AI system interacts with hardware or software that is not part of the AI system itself, the versions of relevant software or firmware, and a description of the forms in which the AI system is placed on the market or put into service (e.g., software package, API, SaaS). This is typically the first document created for any AI system. |
| Design Specification | §2 | A detailed description of the elements of the AI system and the process for its development, including the methods and steps performed for the development, including the use of pre-trained systems or tools, design specifications of the system (the general logic, algorithms, key design choices, assumptions, classification choices), the computational resources used, and the architecture of the system. |
| Development Process | §3 | Documentation of the development methodology, including training techniques and approaches, data requirements, information about training data (provenance, scope, characteristics), data preparation processes (annotation, labelling, cleaning, enrichment), and the selection of parameters and optimisation techniques. This document should demonstrate that a rigorous, documented development process was followed. |
| Monitoring & Testing | §4 | Documentation of the monitoring, testing, and validation procedures used before, during, and after development. Includes test metrics, test results, validation datasets used, testing methodologies (unit testing, integration testing, system testing, acceptance testing), and any identified performance limitations or boundary conditions. Testing documentation is essential for demonstrating Art. 15 accuracy and robustness compliance. |
| Risk Management | §5 | Documentation of the risk management system as required by Art. 9. Includes risk identification, risk analysis, risk evaluation, risk treatment measures, residual risk assessment, and the risk management process lifecycle. This document should reference the specific risks identified for the AI system and the mitigation measures implemented. |
| Data Governance | §6 | Documentation of data governance and management practices as required by Art. 10. Includes data collection processes, data sources, data preparation methods, data quality measures, bias detection and mitigation procedures, personal data handling, and GDPR compliance measures. This document should cross-reference the datasets registered in the Data Governance module. |
| Human Oversight | §7 | Documentation of human oversight measures as required by Art. 14. Includes the description of oversight mechanisms (human-in-the-loop, human-on-the-loop, human-in-command), the competency requirements for overseers, the override and intervention capabilities, and the procedures for escalation. This document should cross-reference the oversight measures registered in the Human Oversight module. |
| Accuracy & Robustness | §8 | Documentation of accuracy, robustness, and cybersecurity measures as required by Art. 15. Includes declared accuracy levels, robustness testing results, adversarial attack resilience measures, cybersecurity measures, and fail-safe mechanisms. This document provides the technical evidence that the system meets the performance and security requirements of the regulation. |
| General | N/A | A catch-all category for supplementary technical documentation that does not fit neatly into the other categories. Examples include deployment guides, integration documentation, user manuals for deployers, change management records, or any other supporting material that contributes to the overall documentation package. |
Creating a New Document
Click + Add Document to open the creation form. Complete all required fields:
| Field | Type | Required | Description |
|---|---|---|---|
| Title | Text input | Required | A descriptive title for the document (e.g., "Fraud Detection Model — System Description v2.1"). Maximum 300 characters. The title should be specific enough to identify both the AI system and the document type at a glance. Titles must be unique within the same AI system — you cannot have two documents with identical titles linked to the same system. |
| AI System | Dropdown | Required | Select the AI system this document pertains to from the dropdown list. The dropdown shows all AI systems in your inventory with their status. If you navigated to the creation form from an AI system's detail page, this field is pre-populated. Each document must be linked to exactly one AI system. |
| Type | Dropdown | Optional | Select the document type from the nine categories described above. Choosing the correct type ensures proper categorisation and helps with completeness tracking — the system can report which Annex IV sections are covered for each AI system. |
| Version | Text input | Optional | The version identifier for this document (e.g., "1.0", "2.3-draft", "2024-Q4-final"). Version tracking is essential for demonstrating that documentation is kept up to date as required by Art. 11. When you create a new version, consider archiving the previous version rather than overwriting it. |
| Content | Rich text editor | Optional | The main body of the document. The rich text editor supports headings, bold/italic formatting, bullet and numbered lists, tables, code blocks, and links. You can draft the content directly in the editor or paste content from external sources. There is no character limit, but extremely large documents (>500KB) may experience slower loading. For documents that already exist in external systems, you can provide a summary and link to the external source. |
Annex IV Completeness Tracking: The system automatically tracks which Annex IV sections are covered for each AI system based on the document types you create. On each AI system's detail page, a completeness indicator shows which of the eight Annex IV sections have at least one approved document. This makes it easy to identify documentation gaps before a conformity assessment.
Document Detail Page
Clicking a document in the list view opens its detail page, which includes:
- Header Section — Shows the document title, linked AI system (clickable link), type badge, status badge, version, and timestamps (created, last updated).
- Completion Percentage — A progress bar showing the document's completion percentage (0–100%). This is a manually set value that you update as you develop the document. The bar is colour-coded: green for ≥80%, amber for ≥50%, red for <50%. The completion percentage is aggregated at the AI system level to provide an overall documentation readiness score.
- Approval Status — Shows who approved the document (if applicable), when it was approved, and the current review status. Documents in "In Review" status display the reviewer's name and the review submission date.
- Content Viewer — Renders the document content in a clean, readable format. The viewer supports all formatting from the rich text editor and provides a print-friendly layout. A "Copy to Clipboard" button allows easy extraction of content for external use.
Editing a Document
From the detail page, click Edit to modify any field. The edit form is pre-populated with current values. Key editable fields include:
- Status — Change from Draft to In Review to Approved to Archived as the document progresses through the review workflow. Status changes are logged in the audit trail.
- Completion — Update the completion percentage as sections of the document are completed. This is a slider or numeric input from 0 to 100.
- Version — Update the version when making significant changes. Consider creating a new document record for major version changes rather than overwriting the existing record.
- Content — Edit the document body using the rich text editor. All content changes are tracked in the audit log.
Tip — Document Templates: When creating your first set of technical documentation, consider establishing templates for each document type. Create a "template" document for each Annex IV section with the required headings and placeholder text, then duplicate and customise it for each AI system. This ensures consistency across your documentation and makes it easier for reviewers and auditors to find the information they need.
Warning — Keeping Documentation Current: Art. 11 requires that technical documentation be kept up to date. If an AI system undergoes significant changes (new version, changed intended purpose, updated training data, modified algorithms), the corresponding technical documentation must be updated. Set reminders to review documentation whenever the linked AI system's version field changes. Outdated documentation is a common audit finding.