ABOUT

Frédéric Verhelst

I work with boards and executive teams on the governance of trustworthy agentic AI — the meaning layer agents reason from, the mandate that governs them, and the liability that comes with it.

Frédéric Verhelst
Education
PhD Applied Physics, TU Delft · MSc Mining Engineering, KU Leuven
Experience
25 years — data, AI and regulated industrial operations
Available for
Non-executive director · Board advisory

I'm fluent in the technology, but my work is strategic, not architectural: I translate what agentic AI can and cannot do into the decisions boards and executive teams must own — the governance, the liability, and the mandate, not the build.

Twenty-five years across data, AI, and regulated industrial operations. As external executive advisor I wrote the digital strategy and value case for the Tyra Redevelopment, a USD 3.35 billion project and the largest single capital investment on the Danish Continental Shelf. I joined Mærsk Oil as Digitalisation Advisor in September 2017 and presented the case at the Final Investment Decision that December, to the partners Shell, Chevron and Nordsøfonden.

I stayed with the Tyra project through the Mærsk-to-TotalEnergies transition, then became Head of Data Office at TotalEnergies EP Danmark, owning the enterprise data mandate. As part-time CEO of the Knowledge Graph Alliance (2024–2025, seconded by TotalEnergies) I led the non-profit bringing industry, academia, and standards bodies together around semantic knowledge graphs. I then moved into product ownership in safety-critical operations.

From August 2026 I am Principal, Energy and Infrastructure at Graphwise, a knowledge graph and semantic AI vendor. I advise boards on agentic AI liability and insurability, and work with C-suites on the Chief Data Officer mandate that The Ontology Imperative sets out. I am available for non-executive director and board advisory appointments focused on corporate AI governance.

The Ontology Imperative is written in a personal capacity. It argues at the level of standards, capabilities and governance structure, and does not assess named vendors in the knowledge graph and semantic technology category. Earlier work that did assess them is marked and frozen rather than withdrawn, and the Reference Library carries the detail.

Board advisory
Agentic AI liability and insurability, in language the board owns.
The CDO mandate
Who governs agentic AI, and how — the office and its operating modes.
The meaning layer
The governed ontology agents reason from — owned, not rented.

For board advisory, speaking, or executive conversations on agentic AI governance and the meaning layer.