Topic
AI + Operations
AI is an operating shift. Most companies are absorbing it without noticing. The structural changes show up in who asks whom for what, which meetings still carry a decision, which layers of the org chart quietly thin out. That's where the actual change happens, and the writing about AI at work rarely gets there.
Start here
-
What AI does to an executive’s read on their own company
Executives stay in touch with their companies by doing the first pass on raw signal themselves. When the model does that pass, the read goes stale faster than anyone notices.
-
AI reorganizes before the org chart does
AI redraws the working chart before anyone redraws the formal one. Leadership notices last.
-
How AI changes the traffic of questions inside a company
Senior people used to stay in touch with the business through a running traffic of small questions, and the traffic is quietly moving into models.
Sub-threads
All memos
-
The GCC has an operating tempo advantage in AI
Every company can buy the same models. The regional advantage is the speed at which leadership can align capital, policy and execution.
-
AI changes capacity before it changes headcount
The first effect is not fewer people. It is uneven spare capacity, hidden queues and a new argument about what the team should do next.
-
AI value does not exist until the operating budget moves
Hours saved and outputs produced are evidence of activity. Value appears when capacity, cost, revenue, risk or service levels change.
-
The CEO’s role in AI transformation starts where the org chart stops
The difficult workflows cross functions, budgets and risk owners. Somebody has to settle the boundary.
-
An AI transformation has ninety days to become an operating system
If the first quarter ends with a roadmap and a use case list, the mandate has not installed anything the business can run.
-
Enterprise AI is an integration program wearing a technology label
The model is the visible part. Identity, data access, systems, handoffs, exceptions and auditability are the work.
-
The AI transformation office should be designed to disappear
A central team can start the work. If it stays the permanent owner, the business doesn't learn to run it.
-
Before you buy AI services, ask what operating layer you’re handing over
Two leading AI vendors moved into the enterprise services layer in May. The procurement question for operators is now which parts of the operating fabric are worth handing to a vendor backed services entity.
-
What boards should ask after the UAE’s agentic government bet
The UAE just moved the benchmark on delegation, and most board packs still can't show what their company has already authorized machines to do.
-
How to actually start
Don't try to automate the company at once. Pick one of the five workflows, build it, and hold the four lines.
-
Why useful AI workflows get clogged
A reader who has built two or three of the workflows in this series will recognize their own day in this essay. Five workflows is one operator. A company is a different problem.
-
Before you approve the AI tool, red-team the operating assumptions
A red-team workflow for executives reviewing AI vendor pitches and internal AI initiatives. The workflow compresses reading time. The decision is always the executive's.
-
Turning messy updates into a weekly operating review
A weekly operating review draft that compresses three hours of synthesis into fifteen minutes of editing. Same questions, every week, in the shape your leadership group already reads.
-
Reducing bookkeeping typing without removing review
A weekly intake assistant that turns a pile of receipts and supplier invoices into a structured intake sheet your accountant opens on Wednesday. The workflow drafts; the accountant posts.
-
The inbox is an operating surface
A daily review queue for an owner-operator running their business across three or four channels at once. Drafts replies, classifies messages, and waits for a person before anything ships.