Good morning. This week we are covering the operating model change, Microsoft cutting 4,800 roles while the model changes underneath, the AI BVF re-run of the agentic HR initiative from issue 13 with the regulatory lens switched on, and a radar where trust in AI agents has reached the standards agenda.
The International Telecommunication Union launched work this week on agent identity, accountability and meaningful human control, because agents are becoming operating actors and governance has to move into the operating model and design.
ITU newsroom, 9 July 2026DigiCert reports AI-related security incidents or vulnerabilities in 78% of organisations surveyed, alongside incomplete inventory and traceability, and a board cannot govern agents it cannot see.
DigiCert, 7 July 2026Google Cloud reports that 83% of organisations need material infrastructure change to capture the agentic opportunity, because data access, integration, compute cost and control architecture decide what reaches scale.
Google Cloud, 2026 State of infrastructure in the agentic AI era, 8 July 2026This week Microsoft announced it is eliminating around 2.1% of its global workforce, about 4,800 people, as part of its company transformation.
Amy Coleman, Microsoft's Chief People Officer, said those roles are not being replaced by AI, and I believe that statement, while also believing the operating model is changing underneath it.
AI changes the volume of coordination, the decisions a manager sees first, the work that moves between teams, and the judgement leaders are expected to make when an agent is now part of the workflow.
A question came to mind: which work has genuinely changed, what decisions have moved, who owns the outcome once an agent is involved, and where does human judgement still have authority to intervene, challenge or stop the work?
This is where many AI programmes are heading, because now the operating model is starting to catch up.
Using the same agentic HR initiative from issue 13, I uploaded through the AI BVF, this time with the regulatory lens switched on.
Result, came back as Fix, leaning towards Stop, held by governance risk, because nothing should move until this was fixed.
The plan it returned was a pre deployment sequence:
It also calculates your waiting time, about EUR 5m to EUR 9m of organisational drag over a 10 week window, remember, that cost still applies even with your deadline pushed out.
Once again, I appreciate your interest and time to read this brief, and as always, look forward to your feedback and comments. Enjoy your weekend, Craig
The Transformation Brief · brief.craighortonadvisory.com · Archive
"We are not replacing people with AI" may be true at the level of a role, but I am seeing enterprises change the work around the role: who prepares the decision, who reviews it, what evidence remains, and who carries accountability when the output is wrong.
That is why the redesign has to come before the project!