Latest thinking
Ideas on AI, leadership, the future of work, and the human dimensions of technological change.
Showing 114 articles
When the AI can do the craft you built your name on, who are you at work?
When AI can do the skill someone spent years mastering, the deeper threat is not redundancy but identity. Since work is a top source of meaning, protecting it becomes a leadership task. How to help people move their sense of worth to the work a machine cannot do.
Your training budget clears this year's skills gap. Next year's is already forming.
The WEF finds 39% of core skills will change by 2030 and 59 in 100 workers need reskilling. A one-off course clears one wave; the gap keeps reforming. Why reskilling has to become a standing capability, built into the work.
Your AI vendor says the model might be conscious. Is that a board problem, or sci-fi?
In its February 2026 safety report, Anthropic's model put its own odds of consciousness at 15 to 20 percent. Not proof of anything, but now in official documentation. Why the board question is not whether it is conscious, but what your exposure is if others treat it as if it might be.
Your whole AI stack now runs on one vendor. What happens when you want to leave?
Standardising on one AI provider is fast to start and expensive to leave, because lock-in now stacks across model, orchestration, data, governance and know-how. The most capable buyers are already hedging with more than one provider. How to design portability in and keep your bargaining power.
You gave your team the same AI everyone raves about. Why did the results come back flat?
The same AI can raise performance or sink it below no AI at all, depending entirely on how people and the tool are arranged to work together. Why the value lives in the arrangement, not the machine, and how to design one where human judgement compounds rather than quietly wastes away.
You gave everyone the same AI. Why did it make your best people pull further ahead?
Anthropic's usage data shows skilled users pull far more from AI than beginners, so a general tool handed out flat tends to widen the capability gap, not close it. But who gets lifted is a design choice. How to deploy AI so it raises the whole team, not just the top.
Your D&O insurance is starting to exclude AI. Directors are personally exposed.
In 2026, regulators increasingly expect directors to evidence AI oversight, just as insurers begin writing AI exclusions into liability and D&O cover. More personal responsibility, a thinning safety net. The concrete steps a board should take before a hard question arrives.
AI makes each of your people more creative. It quietly makes all of them more alike.
A 2024 Science Advances study found generative AI made individual writers more creative, but their work more similar to each other. AI lifts the floor and flattens the peaks, trading distinctiveness for polish. Use it to raise your team, not converge them into the crowd.
AI keeps winning you the quarter. Is any of it building the decade?
Almost every AI business case is a short-term efficiency: cost down, speed up. A review of 1,377 studies argues AI's deeper promise is escaping short-termism, better resource allocation, real stakeholder engagement, and value that lasts. Point it at the long game.
AI just made software cheap to build. Your build-versus-buy answer may have flipped.
Building software was slow and costly, so the safe default was to buy. AI has moved that line: generative coding is a 2026 breakthrough, and the cost of producing code is collapsing. Re-run build-versus-buy, but remember cheap to build is not free to own.
The headlines say AI is coming for the jobs. The usage data says something calmer.
Predictions about which jobs AI will erase are cheap. Real usage data tells a steadier story: mostly augmentation, spreading task by task rather than job by job. How to plan a workforce from evidence instead of headlines.
You built AI controls for a tool that advises. Now it acts. Do they still hold?
Most organisations matured sensible controls for AI that advises: inaccuracy, cybersecurity, IP. Agents act, which changes the shape of the risk. McKinsey's 2025 data shows controls catching up to yesterday's risks as the surface shifts.