NLEN

DEEP DIVE · DUTCH WOO

AI in the Woo process: support without taking over responsibility.

A public-information request is not a single search. It is an administrative and legal process in which intake, scope, searching, assessment, redaction, review and decisions must be demonstrably careful.

AI advises, structures and checks. Authorised staff investigate, assess and decide.

The opportunity is targeted support at each step: finding relevant sources, structuring results, flagging potentially sensitive passages and checking consistency. Context, legal interpretation and formal responsibility remain human.

This is a general process exploration, not legal advice. Specific implementation depends on the organisation, policy, systems and request.

PHASE 1 · STEPS 1–3

Intake, qualification and treatment strategy

Register the request, assess whether the Woo applies and define ownership, planning and communication. AI can structure text, flag dates and suggest clarification questions; staff determine the formal route.

1. Receive and register
Capture channel, date, requester, requested information, relevant units and deadlines consistently.
2. Initial Woo qualification
Assess jurisdiction, existing public information and whether clarification or forwarding is needed.
3. Determine treatment strategy
Set ownership, stakeholders, planning, search approach and escalation routes.

PHASE 2 · STEPS 4–5

Process analysis and search plan

Reconstruct where information was created, changed and stored. AI can propose entities, periods, synonyms and source locations; staff establish scope.

4. Process and file analysis
Map decisions, people, systems, mailboxes, case files and external parties.
5. Create the search plan
Document sources, terms, periods, owners, exceptions, logging and completion criteria.

PHASE 3 · STEPS 6–7

Search, collect and record results

AI can support semantic search, grouping and classification. Staff coordinate source holders, investigate gaps and explain why the search was sufficient.

6. Conduct the search
Search agreed sources reproducibly while preserving context, relationships and integrity.
7. Record results
Record what was found, where, how and by whom, including empty results and limitations.

PHASE 4 · STEPS 8–10

Assess disclosure, redact and involve third parties

AI can flag personal data, similar assessments and inconsistent redactions. Staff apply legal grounds, weigh interests and make formal judgements.

8. Assess disclosure
Review each document or passage in context; never apply exceptions mechanically.
9. Anonymise and redact
Check text layers, metadata, attachments, reversibility and accessibility.
10. Third-party review
Identify stakeholders and process views where required; the formal balance remains human.

PHASE 5 · STEPS 11–14

Decision, quality control, disclosure and learning

AI can structure drafts and flag inconsistencies. Staff establish the decision, verify reasoning, deliver documents and archive the file.

11. Draft the Woo decision
Connect scope, search accountability, inventory, exceptions and reasoning.
12. Legal quality review
Check authority, procedure, reasoning, consistency, deadlines and document alignment.
13. Decision and disclosure
Send securely and verify readability, accessibility, integrity and redaction.
14. Evaluation and archiving
Archive the treatment file and use lessons to improve information management.

Conditions for reliable support

AI is useful only when the organisation controls data, sources, access, quality and human responsibility.

TraceabilityBack to source and passage.
Protected processingAppropriate hosting and authorisation.
Human reviewNo autonomous formal decisions.
Testing and loggingMeasure, record and improve.

BEACON IN CONTEXT

A practical application

Beacon supports municipalities in finding, collecting and preparing information with locally deployable AI and source traceability. It does not replace the Woo officer or make formal disclosure decisions.

Visit Beacon →