How WorkWithAI content is created and reviewed
A transparent account of authorship, evidence, translation, participant proof and the limits of AI-assisted publishing.
Last reviewed: 22 July 2026
Authorship and human review
Arnaud Llamas Bravo authors the workshop programmes and remains the accountable human reviewer for every published page. Training material is checked against the stated learning outcomes, audience and delivery format before publication.
Sources and fact-checking
Product, privacy and regulatory claims are checked against current primary sources. Microsoft Learn and Microsoft Support are used for Microsoft 365 Copilot, Copilot Studio and Power Automate. Belgian Data Protection Authority and European Commission materials are used for privacy and governance.
Microsoft product-change reviews
Microsoft products change frequently. Product-dependent pages are reviewed after material Microsoft announcements and during scheduled editorial checks. Features, availability and licences must still be confirmed for each organisation before delivery.
Translation review
English, French and Dutch pages are reviewed for meaning, business terminology and Belgian usage. Localized pages preserve the same core programme and safety boundaries; they are not literal machine translations.
Testimonials and training-count methodology
The public testimonial source is testimonial.to. The supplied private export contains 247 unique testimonial records dated from 13 October 2023 to 8 July 2026; private fields and email addresses are never published. The figure ‘8,500+ professionals trained since January 2024’ is a founder-attested cumulative count, last reviewed on 22 July 2026, and is not presented as independently audited.
Organisation references
References to Proximus, Honda and the Flemish Government are founder-confirmed and authorized examples of organisations served. They do not imply a partnership, endorsement or certification.
AI-assistance disclosure
AI tools may assist with drafting, translation comparison, code, metadata and quality checks. Arnaud reviews the public meaning, factual claims and professional boundaries. AI output is never treated as an authoritative source.
Corrections and updates
Corrections are prioritized when a source changes, a product claim becomes outdated or a reader flags an error. Send the page URL and proposed correction to arnaud@workwithai.be. Material corrections are reviewed by the responsible editor.