Executive Bytes · the newsletter
A weekly newsletter by Dr. Mechie Nkengla, CEO of Data Products. Each edition takes one shift in AI, data or technology and turns it into the questions a leadership team should be asking. Read it on LinkedIn, or in full here, with its sources.
The latest edition
November 9, 2026 · Governance and trust
AI agents are getting access, responsibilities and budgets. What they may do, what people must still own, and how to tell whether any of it is paying off.
In one line
The agent may never ask for a performance review. The business should insist on one anyway.
Every edition
Each edition is written up in full here, with the sources it names and the questions to take into your own meetings.
Governance and trust
AI agents are getting access, responsibilities and budgets. What they may do, what people must still own, and how to tell whether any of it is paying off.
On LinkedInStrategy and delivery
Q4 is where transformation plans meet the calendar. How to tell the projects that can still deliver from those producing activity and expensive scenery.
On LinkedInAI economics · From the table
From a Private Executive Bites table in Chicago: AI cost is measured at the meter, but the unit that matters is the quality-adjusted business outcome.
On LinkedInGovernance and trust
When content is made and altered at machine speed, trust cannot rest on appearance. Leaders need a verifiable record of origin, AI use and approval.
On LinkedInPeople and literacy
Companies are buying AI faster than they build the judgment to use it. The advantage is knowing what to delegate, what to verify, and when to stop.
On LinkedInSecurity
A familiar voice, a convincing video or a complete identity record can now be faked cheaply. Verify authority through the process, not the performance.
On LinkedInAI economics
AI arrives on a software invoice. At scale it behaves like capacity rights, multi-year commitments and long-duration bets that need capital discipline.
On LinkedInAI economics
AI has moved beyond the software budget. The next constraint is physical: compute, memory, networks, data centers, power and the capital to secure them.
On LinkedInAgentic AI
Agents arrive as features, connectors and staff-built automations, not programs. How to find the ones already at work in your business, and govern them.
Written for the web
Governance and trust
Most gen AI policies were written for a chatbot in a browser tab. Where they now fall behind, and five steps to put governance where AI actually runs.
Written for the web
Policy and geopolitics
AI nationalism: governments are building sovereign AI, tightening chip exports and embedding language models into propaganda, defense and diplomacy.
On LinkedInPolicy and geopolitics
Who owns the moon? The race for lunar data, resources and AI supremacy, and why the fight will be over telemetry, algorithms and data rights.
On LinkedInPeople and work
AI layoffs 2.0: as AI-driven cuts reach middle managers and HR, who is left to mentor, resolve and remember? What leaders should protect now.
On LinkedInTrust and security
Fighting AI disinformation with AI: detection tools, a dismantled bot farm, why people miss deepfakes, new liability rules, and how to cut your exposure.
On LinkedInTrust and security
Deepfakes, bad bots and fading trust are now brand risks. Part 1 maps the threat to your brand and the first moves: detection, partners and a drill.
On LinkedInAgentic AI
Agentic AI and people: retail workers' job fears, a patchwork of U.S. AI rules, and financial services in 2030. Part 2, with three moves to make now.
On LinkedInAgentic AI
Agentic AI acts without waiting for a prompt. Part 1 weighs its promise against its perils: productivity, competition, failure risk and customer trust.
On LinkedInData teams
Why data and software teams should work as one: a data-driven culture, five practical steps to collaborate, and the data integrity gaps to close first.
On LinkedInAI in operations
Operationalizing AI takes more than a model: a defined problem, governed data, trained people, pilots and ethical guidelines. A guide for executives.
On LinkedInMachine learning
TinyML puts machine learning on microcontrollers and wearables: faster decisions, lower power, better privacy, and the limits to weigh before you start.
On LinkedInRegulation
The EU AI Act, a fragmented U.S. approach and new national laws: what the coming wave of AI legislation means for companies deploying AI across borders.
On LinkedInData platforms
Hybrid cloud gives AI the storage, compute and security it needs. Three pieces on aligning cloud with goals, placing workloads and staying vendor-neutral.
On LinkedInHow an edition works
1
Each edition goes to more than 1,400 subscribers on LinkedIn: one shift, the evidence, and the boardroom questions it raises.
2
The same edition with its sources linked, the areas to watch numbered and drawn, and links to the work behind each point.
3
Each page links to its LinkedIn edition for the discussion, and each LinkedIn edition links back here. A few pieces are written for the web only.
From the table
Some editions start at a Private Executive Bites table, a small dinner of senior leaders around one hard question. AI cost is being measured in the wrong place came from the Chicago table in September.
About Private Executive BitesTell us the question. It may become a future edition, or a 30-minute conversation about your own case.
Read Executive Bytes on LinkedIn? Suggest the next topic; it may become an edition.
