ApexTransform Insights & Selected Publications

Ideas for decisions. Insights for scale.

Analysis by Stéphane Gervais on trusted AI, industrial transformation, cybersecurity and technology scale-up.

Browse the collection

All insights and publications

1–12 of 84 publications

Industrial controller with exposed software layers and vulnerability warnings, illustrating the Cyber Resilience Act reporting deadline in September 2026 and full product requirements in December 2027.
Cybersecurity & Human Trust

The CRA Is Already Here: Why 2027 Is the Wrong Deadline to Watch

· 11 minutes

The EU Cyber Resilience Act's main product requirements apply in December 2027, but mandatory reporting began in September 2026. Crucially, the obligation also covers products already on the market and can extend beyond their support period. For manufacturers, the immediate challenge is not certification. It is maintaining enough knowledge of deployed products, third-party dependencies and security events to make def

ApexTransform Insight
Read the insight
Shane Buckley, President and CEO of Gigamon. Illustration for ApexTransform Insights. Credit: Portrait supplied by Gigamon/Hotwire; editorial visual treatment by ApexTransform.
Cybersecurity & Human Trust

The Next AI Security Failure May Look Completely Legitimate

· 6 minutes

AI security failures may look entirely legitimate: a valid identity, approved application and authorised network path can still enable a harmful action. Shane Buckley argues that machine-speed attacks require faster detection, but visibility alone cannot prove intent. Enterprises must define authority before AI agents act—automating only narrow, reversible responses and reserving consequential actions for pre-approve

ApexTransform Insight
Read the insight
Conceptual illustration of an autonomous AI agent, independent security enforcement boundary and enterprise computing infrastructure.
Physical & Industrial AI

NVIDIA Is Moving AI Agent Security Into the Infrastructure Layer. The Next Question Is Authority.

· 14 minutes

At NVIDIA GTC Berlin 2026, the new OpenShell architecture brings independent enforcement and hardware-assisted protection to autonomous AI agents. Stéphane Gervais examines what these controls achieve, where their guarantees stop and why technical permission must be connected to legitimate human authority.

ApexTransform Insight
Read the insight
Brian Comiskey presenting the “Age of Intelligence” at FDDay in Paris, illustrating the shift toward vertical AI and agentic AI ahead of CES 2027.
Physical & Industrial AI

From Assistant to Actor: Building AI in the Absence of Trust

· 8 minutes

Brian Comiskey, Vice President of Innovation & Trends at the Consumer Technology Association, expects CES 2027 to show AI moving from assistant to actor, across multi-agent software, specialized models and robots. His answer to the trust problem comes from cybersecurity rather than from AI, and it points to a practical discipline for any organization delegating work to machines.

ApexTransform Insight · Originally on ApexTransform Insights
Read the insight

The featured Insight above is included in the 12 entries on this page.

From analysis to action

Turn an insight into an executive decision.

Use ApexTransform's decision tools to clarify the board mandate, quantify one priority AI use case and define the next 90 days.