Resources
The blog
Perspectives, field feedback and pedagogy: to think about useful AI, with humans at the centre.
Perspectives, field feedback and pedagogy: to think about useful AI, with humans at the centre.
An Asma Mhalla talk at a MEDEF event first annoyed me, then took me a week to digest. What I take from it: on the ground, it isn't the attack that destroys the most value, it's the fear that freezes everything.
Read the article →The gadget, the uncontrolled hallucination, the ritual no one keeps alive, the managerial problem disguised as tech: where AI will fix nothing.
Read the article →Match carrier invoices against real services and weights, rebill the right customer, and finally see clearly: AI roughs out, the human arbitrates.
Read the article →Andrej Karpathy says it plainly: "vibe coding" is giving way to agent engineering. A signal that reaches far beyond code, and speaks to any AI-driven transformation.
Read the article →In EHS, AI prepares and structures the signatory's work, but the decision that carries legal responsibility stays strictly human.
Read the article →In a regulated environment, AI can absorb the formatting and leave the expert on the substance, without ever touching the decision or the batch release.
Read the article →Your sales reps answer the same customer questions ten times a week: an AI knowledge base can give that time back.
Read the article →Leaks, hot spots, tamper-proof readings: inspection robots have real uses. But sometimes a simple sensor does the job, and it has to be said.
Read the article →One or two week forecasts rarely land right. AI roughs out a number, the planner adjusts with field knowledge.
Read the article →A written work instruction doesn't transmit the gesture. How to film know-how and turn it, with AI, into training newcomers actually understand.
Read the article →A full MES often exceeds an SME's budget, but AI can consolidate your OEE readings and turn them into a real dashboard.
Read the article →Before buying sensors, make visible the data you already capture. It's often the fastest gain, and the cheapest.
Read the article →There is no best AI in the abstract, only the right AI for the right use, chosen for your context.
Read the article →The real question isn't "agents versus software," but intelligence as a system: orchestrating rather than replacing.
Read the article →Anticipate rather than endure: predictive AI, paired with human vigilance, builds safer and higher-performing sites.
Read the article →The fastest and least risky AI gains live in daily admin work: meeting minutes, emails, reporting.
Read the article →As in elite sport, AI-driven transformation is first and foremost a matter of culture, preparation and consistency.
Read the article →Well placed, AI eases cognitive overload and returns attention to what matters: a support, not a substitute.
Read the article →The best way to start an AI project isn't to pick a tool, it's to spend a day on the shop floor.
Read the article →Industrial waste has changed in nature: it is no longer measured only in lost time, but in misplaced intelligence.
Read the article →After twenty-five years running plants, I created Angle Intelligence. Not to ride the AI hype, to do exactly the opposite.
Read the article →