Bertrand Duperrin @duperrin.com · 07/09/2026Been recently asked what the perfect profile for an Chief AI officer is. IMO no perfect profile, just depends on the problem you want to solve and the bottlenecks you're facing. A long post made short with an image (yes IA is great when you have ideas and absolutely zero design skills). 000
Bertrand Duperrin @duperrin.com · 08/07/2026Weekly wrap-up #5 : Economic scrutiny, organizational transformation and workforce adaptation redefine AI’s next phase as sustainable execution overtakes technological performance as the decisive competitive advantage.bduperrin.substack.com2025.28: the challenge of dual organizations, the AI economy is sputtering, society is worried...Economic scrutiny, organizational transformation and workforce adaptation redefine AI’s next phase as sustainable execution overtakes technological performance as the decisive competitive advantage. 000
Bertrand Duperrin @duperrin.com · 01/07/2026Weekly wrap-up #4 : AI spending pressures, implementation challenges and governance tradeoffs reveal that long-term value depends less on adoption than on disciplined execution and organizational change.bduperrin.substack.com2025.27: ATS, AI, and legal risks in recruitment, AI budget time bombs, AI agents and implicit rules...AI spending pressures, implementation challenges and governance tradeoffs reveal that long-term value depends less on adoption than on disciplined execution and organizational change. 100
Bertrand Duperrin @duperrin.com · 22/06/2026Weekly wrap-up #3 : Why AI adoption is outpacing value creation, forcing organizations to rethink work, governance, software capabilities and the economics of automation.substack.com2026.26 : The dilemma of laying off or growing, AI & operations, Meta's mistakes...Why AI adoption is outpacing value creation, forcing organizations to rethink work, governance, software capabilities and the economics of automation. 000
Bertrand Duperrin @duperrin.com · 18/06/2026Weekly wrap-up #2 : Why AI’s future depends less on better models than on organizational design, cost discipline, governance, and the ability to continuously adapt.bduperrin.substack.com2026.25 : Management hidden's face, the risks of automation, AI's risky funding...Why AI’s future depends less on better models than on organizational design, cost discipline, governance, and the ability to continuously adapt. 000
Bertrand Duperrin @duperrin.com · 12/06/2026Btw my weekly wrap-up (issue #1) : as AI moves beyond model performance, economics, infrastructure, organizational design, ROI pressures and public skepticism increasingly shape adoption decisions.substack.com2026.24 : AI, costs, infrastructure, employment, and work organizationAs AI moves beyond model performance, economics, infrastructure, organizational design, ROI pressures and public skepticism increasingly shape adoption decisions. 000
Bertrand Duperrin @duperrin.com · 12/06/2026Trying to find my way back to bluesky. Not easy to find the right balance between many platforms that matter to me for some reason. I'll possibly share my thoughts and concerns later on since I think many of you have been through the same process. 020
Bertrand Duperrin @duperrin.com · 31/12/2025AI First unfolds in three stages: acceleration (efficiency), extension (new capabilities), and transformation (redesign of roles and flows). Only by reaching this third level can AI shift from tool to design principle and deliver true organizational impact.duperrin.comThe promise of AI: acceleration, expansion, transformationArtificial intelligence is often presented as a driver of efficiency, a means of working faster or reducing the burden of repetitive tasks. This perception is w 010
Bertrand Duperrin @duperrin.com · 31/12/2025The real AI challenge for businesses isn’t chasing model upgrades, but building “programmable” organizations. Agentic AI needs clear goals, structured knowledge, and local context. True impact requires managerial clarity, strong infrastructure, and a political response to shifting work dynamics.academy.shiftbase.infoCan AI Help Reverse the Oversimplification of Management?For leaders to be programmers of their organisational OS, they need AI's help in embracing complexity and reality, rather than relying on simplistic data proxies 000
Bertrand Duperrin @duperrin.com · 30/12/2025AI often impresses but rarely convinces, because users don’t need wow, they need utility. Flashy demos don’t translate into real gains if everyday tasks aren’t improved. What matters isn’t how far the model goes, but how much faster it helps where time is actually lost.theintelligencefabric.comAI is Impressive. So Why Does No One Care?Silicon Valley is obsessed with engineering “wow” moments. But for the rest of the world, a “super-smart” AI that can’t handle basic arithmetic isn’t a miracle, it’s a distraction 010
Bertrand Duperrin @duperrin.com · 30/12/2025AI only transforms organizations when teams rethink how they work. Without structural change and shared goals, local gains remain isolated. Real impact comes from collective adoption, role redefinition, and embedding AI into a coordinated, team-driven transformation. www.duperrin.com/english/2025...duperrin.comCollective appropriation of AI: the only condition for tangible impactBusiness still tend to associate an employee's use of a tool with an increase in their performance or even a transformation in their working methods. They are d 000
Bertrand Duperrin @duperrin.com · 30/12/2025AI only boosts productivity once a structural transformation threshold is crossed. Without rethinking workflows and building "architectural literacy", its use remains superficial. Productivity gains don’t come from access alone, but from deep integration into how work is organized.dataconomy.com“The LLM productivity cliff”: New research offering a different lens on AI productivityThe dominant story about large language models is simple: give workers AI, and productivity rises. But workplace evidence is uneven. 010
Bertrand Duperrin @duperrin.com · 30/12/2025Fear in meetings signals deeper dysfunction. When people hold back, group effectiveness drops. Psychological safety, not control, drives performance. Inclusive leadership, where mistakes and questions are welcome, turns meetings into real workspaces, not zones of silence or self-protection.duperrin.comNo one should be afraid to walk into a meetingEveryone has had the experience of walking into a meeting. The door opens, faces are closed, you take your seat, you breathe a little too loudly, some people lo 030
Bertrand Duperrin @duperrin.com · 27/12/2025Pernod Ricard’s digital shift succeeded by fostering demand instead of imposing change. Through local pilots, support, and adjusted expectations, AI tools once seen as intrusive became valued aids. The real transformation was cultural, proving adoption grows through trust, not pressure.hbr.orgHow a French Spirits Company Created Employee Buy-In for AIDigital transformation often stalls when employees resist new technology. To overcome this common challenge, French spirits company Pernod Ricard drew upon four strategies: proving value through A/B t... 000
Bertrand Duperrin @duperrin.com · 27/12/2025“AI First” covers divergent realities : tech investment, strategic vision, organizational change, or experimentation. Without clarity, it risks becoming an empty slogan. Defining its meaning helps turn intention into structured action aligned with each company’s true goals.duperrin.comAI First: four definitions that have nothing to do with each otherSince the concept of “AI First” was born, it has become, if not a real trend in organizational transformation, then at least a real topic of discussion 000
Bertrand Duperrin @duperrin.com · 26/12/2025AI is reshaping jobs unevenly: seniors are let go for tech that’s not ready, while entry roles vanish, blocking young graduates. This misaligned shift, driven by hype over results, weakens internal learning and innovation. A reset is needed to align talent strategy with real AI capabilities.hrexecutive.comThe AI layoff trap: Why half will be quietly rehiredNew research shows entry-level jobs disappearing while experienced workers lose positions to AI that doesn't exist yet. 000
Bertrand Duperrin @duperrin.com · 26/12/2025AI adoption is slow but real, held back by tech maturity and organizational inertia. Impact lies in gradual process improvement, not flashy demos. Early adopters gain agility, but long-term value depends on smart data use and aligning AI with existing workflows, not replacing them outright.mckinsey.com 000
Bertrand Duperrin @duperrin.com · 23/12/2025Gregg Popovich shows that great leadership blends trust, adaptability, and human connection. His style balances rigor with flexibility, tailoring support to individuals while staying rooted in clear principles, proving that authority grows stronger when it listens, adjusts, and stays consistent.duperrin.comLeading means learning to make room for others, an imaginary interview with Gregg PopovichAt a time when businesses are rediscovering the importance of trust, adaptability, and care in leadership, Gregg Popovich's career path offers us some fascinati 010
Bertrand Duperrin @duperrin.com · 22/12/2025Digital twins, once space tech, now model cities, bodies, and systems in real time. By blending AI, simulation, and live data, they test scenarios, cut costs, and boost performance. Their true power lies not in AI alone, but in mastering data engineering at scale.processexcellencenetwork.comSurprising things about digital twinsWhile digital twins are becoming increasingly popular, implemented, and understood, there are still many surprising things about digital twin technology 000
Bertrand Duperrin @duperrin.com · 22/12/2025Automation and AI reshape skills, but short-term productivity hides a deeper risk: the erosion of skill renewal. As junior learning tasks disappear and roles become hybrid, skills act like volatile assets. Without sustained investment, firms fuel a fragile productivity bubble.itsocial.frL’automatisation fragiliserait la capacité de renouvellement des compétences - IT SOCIALL’automatisation et l’IA transforment profondément la structure des compétences dans les entreprises. Mais au-delà des pénuries de talents largement commentées, une fragilité plus profonde se dessine.... 000
Bertrand Duperrin @duperrin.com · 22/12/2025"AI First" means prioritizing AI as a starting point, not excluding humans. Confusing it with "AI Only" leads to blind automation. True AI strategy blends tech with human judgment, aligning tools with intent and structure rather than chasing automation for its own sake.duperrin.comAI First is not AI Only: clarify your intentions before transforming your businessWhen businesses talk about their ambition to become "AI First", there is often confusion about associating this orientation with widespread automation, as if a 000
Bertrand Duperrin @duperrin.com · 21/12/2025In 2025, HR tech saw widespread AI adoption, major consolidation, and ROI-driven choices. Applied and generative AI spread across all HR functions. M&A activity rose, while HR’s role shifted toward strategic impact, moving beyond a cost-center logic.unleash.aiHR technology trends in 2025: AI adoption, ROI focus, and market consolidationUNLEASH speaks to Ralf Hofmann, Managing Partner & Co-Founder at Drake Star to discuss the top trends in the HR technology market over the past 12 months, as well as looking forward in 2026. 000
Bertrand Duperrin @duperrin.com · 21/12/2025When managers shift from quick fixes to active observation, teams self-regulate, tensions ease, and decisions flow better. Inspired by agile thinking, this approach favors timing, clarity, and collective intelligence over control or rigid action plans.qualitystreet.frLe Manager est l'observateur de son systèmeQuand le manager devient un observateur attentif de son système, le système commence à évoluer… dans le bon sens 010
Bertrand Duperrin @duperrin.com · 20/12/2025“AI First” means starting with AI as a design choice, not excluding humans. Confusing it with “AI Only” leads to blind automation. True AI strategy balances tech with human judgment, focusing on structure, intent, and performance, not automation for its own sake.duperrin.comAI First is not AI Only: clarify your intentions before transforming your businessWhen businesses talk about their ambition to become "AI First", there is often confusion about associating this orientation with widespread automation, as if a 000
Bertrand Duperrin @duperrin.com · 19/12/2025In 2025, HR tech saw mass AI adoption, platform consolidation, and ROI-driven decisions. Applied and generative AI spread across all HR functions. M&A surged as firms moved from point solutions to integrated systems. No bubble in sight, just a strategic shift toward efficiency and impact.unleash.aiHR technology trends in 2025: AI adoption, ROI focus, and market consolidationUNLEASH speaks to Ralf Hofmann, Managing Partner & Co-Founder at Drake Star to discuss the top trends in the HR technology market over the past 12 months, as well as looking forward in 2026. 000
Bertrand Duperrin @duperrin.com · 19/12/2025When managers shift from quick intervention to active observation, team dynamics improve. Inspired by agile and lean thinking, they see the organization as a living system. Management becomes less about control, more about timely support, clarity, and fostering collective self-regulation.qualitystreet.frLe Manager est l'observateur de son systèmeQuand le manager devient un observateur attentif de son système, le système commence à évoluer… dans le bon sens 000
Bertrand Duperrin @duperrin.com · 18/12/2025AI in hiring doesn't just reflect fairness but it redefines it. A study shows how algorithmic tools replaced human judgment with standardized metrics, sidelining context and diversity. When fairness is coded into systems, it must stay open to debate, not be mistaken for neutral truth.hbr.orgNew Research on AI and Fairness in HiringAI promises to make hiring fairer by reducing human bias, yet it often reshapes what fairness means—and locks in one definition. A three-year study of a global consumer-goods firm found that their alg... 000
Bertrand Duperrin @duperrin.com · 18/12/2025HR too often acts downstream, fixing damage caused by flawed management systems. Rooted in an administrative legacy and reinforced by managerial disengagement, HR remains reactive. Repositioning HR around work design, integrated with operations, shifts focus from repair to performance.duperrin.comHR is not the management's fire departmentIn many businesses, human resources departments don't intervene at the right time. They arrive when everything is going wrong, when a manager is losing control, 000
Bertrand Duperrin @duperrin.com · 17/12/2025Organizations accumulate invisible debts: technology debt and workforce debt. They don’t appear on balance sheets but erode performance through poor tools, outdated processes, overload, and underinvestment in skills. Short-term fixes create long-term drag.bloorresearch.comYour business is in Debt - and it’s not FinancialMost organisations carry debts they don’t see. Not financial debts, but something far more crippling: Technology debt and Labour debt. 000
Bertrand Duperrin @duperrin.com · 17/12/2025Organizations are complex systems. Linear models fail; value emerges from interactions, feedback loops, and adaptation. Leadership means setting boundaries, observing behaviors, and redesigning the frame, not controlling outcomes or optimizing rigid structures.engagedorgs.comPeople Are Complex - Our Systems Are NotPeople Are Complex but the governance of organizations does not differentiate between their simple, complicated, and complex systems. 000
Bertrand Duperrin @duperrin.com · 17/12/2025Turning AI gains into performance requires clear governance. AI creates productivity, but without strategic intent those gains stay diffuse. Only deliberate choices on how to use saved time, for skills, quality or innovation, turn AI potential into lasting value.duperrin.comWithout governance, the gains from AI are virtualBusinesses often talk about the benefits of AI, particularly in terms of productivity and revenue, as if they were the logical outcome of successfully adopting 000
Bertrand Duperrin @duperrin.com · 17/12/2025Generative AI doesn’t fix organizations, it exposes them. When layered onto broken structures, it amplifies dysfunction. Only firms that redesign work, roles, governance, and incentives see value. The real challenge is managerial and organizational, not technological.thefutureorganization.comHow to Embrace Generative AI Without Automating the Soul of WorkGenerative AI insights from Rebecca Hinds of Glean on AI transformation, workflow automation, and organizational design to protect employee experience. 000
Bertrand Duperrin @duperrin.com · 16/12/2025Companies struggle with AI not for lack of tools, but of work design. Leaders rethink how work really happens, redistribute tasks between humans and AI, redesign roles and decisions, and embed learning. AI works when organization changes first.duperrin.comHow can you prepare your organization for AI?Reading some of the current debates on AI, one might think that everything hinges on the sophistication of models or the speed of tool deployment. But the BCG a 000
Bertrand Duperrin @duperrin.com · 16/12/2025AI often fails to scale because organisations lack an operating model. Not structure or process, but a system aligning intent, capability and value in constant change. AI only creates impact when it makes this model living, mutable and measurable.bloorresearch.comWhy AI Needs an Operating Model and You Don’t Have OneWhy AI Needs an Operating Model and You Don’t Have One here is a growing belief that AI will change everything. 010
Bertrand Duperrin @duperrin.com · 16/12/2025NVIDIA is not Enron, but its model raises concerns. An internal memo sought to dismiss fraud comparisons while addressing fears of circular financing, debt-fuelled GPU demand, underused capacity, and weak AI revenues. The risk is not scandal, but a slow unwind of a leveraged AI value chain.wheresyoured.atNVIDIA Isn't Enron - So What Is It?If you enjoy this free newsletter, why not subscribe to Where's Your Ed At Premium? It's $7 a month or $70 a year, and helps support me putting out these giant free newsletters! At the end of Novembe... 010
Bertrand Duperrin @duperrin.com · 15/12/2025AI reshapes work, but skills still matter. As automation grows, human judgment, empathy and critical thinking gain value. Overreliance on AI risks skill atrophy. HR must invest in early careers and learning, training talent to complement AI rather than compete with it.hrexecutive.comWill skills still matter in 2026?The rise of AI means orgs may face a growing dearth of skills—complicating the already uncertain landscape for early-in-career talent. 000
Bertrand Duperrin @duperrin.com · 15/12/2025Hierarchy emerges naturally in human groups to reduce uncertainty and enable collective action. Research shows that even without rules, people recreate order. Participatory models work only with support from authority. The goal is not to erase hierarchy but to make it flexible, balanced and open.duperrin.comThe horizontality of society is a collective illusionSince the early days of the web, and even more so with the advent of social media, there has been widespread talk of the promise of a horizontal world, fre 000
Bertrand Duperrin @duperrin.com · 15/12/2025Performance management must match the pace of modern work. Static systems and rare feedback don’t fit a world of shifting priorities. LeAnne Lagasse urges HR to bring clarity, alignment and real-time feedback into daily workflows, activating performance through dynamic, human-centered support.betterworks.comDesigning Performance Management Systems That Can Meet the MomentWork is moving faster than traditional reviews can keep up. Discover how HR can use AI and better communication to enable continuous performance and alignment. 000
Bertrand Duperrin @duperrin.com · 15/12/2025AI transformation works only when business leaders can link operations and technology. Few have this hybrid skill today. Developing it through continuous, hands-on learning creates a small group of leaders able to redesign entire domains and drive impact.mckinsey.com 000
Bertrand Duperrin @duperrin.com · 14/12/2025HiPPO shows how leaders’ opinions still override evidence. In uncertainty, authority replaces method, and data becomes justification instead of guidance. The issue is structural, not personal. Only clear roles, valued dissent, and practices like “the boss speaks last” break the cycle.duperrin.comThe HiPPO principle: when opinion replaces decision-makingWe thought we had replaced the power of the boss with highly rational approaches, particularly those based on data. A century of management science, from Taylor 000
Bertrand Duperrin @duperrin.com · 13/12/2025AI doesn’t fail, leaders do. Most projects collapse not from tech limits but from missing clarity, ownership, and aligned decisions. The tech is ready; organizations aren’t. Only committed leadership and real operational change turn AI into impact.dataconomy.comWhy 84% of AI projects fail—and it’s not the technologyYour AI project isn't failing because the models aren't good enough. It's failing because your leadership team is. RAND Corporation's 010
Bertrand Duperrin @duperrin.com · 12/12/2025Overwhelm is a sudden overload that stays hidden behind apparent performance. It arises when control, clarity and recovery collapse, pushing even top performers to the edge. The issue is organizational, not individual, and only redesigned work can prevent silent breakdowns.hbr.orgDo You Know If Your Team Is Overwhelmed?In a world where relentless demands and blurred boundaries have become the norm, overwhelm is no longer an occasional experience; it’s a defining feature of modern work. It’s easy to dismiss it as str... 000
Bertrand Duperrin @duperrin.com · 12/12/2025AI-native firms embed data and learning from day one. Others must retrofit rigid structures and slow processes. Becoming AI First is a design shift, not a tech upgrade, and the real question is how close an organization can realistically get.duperrin.comA business can be born "AI first", but can it become one?The expression "AI First" is now so prevalent in discourse that, in one way or another, it is, if not a goal for businesses, at least a focus for reflection and 000
Bertrand Duperrin @duperrin.com · 12/12/2025AI is widely adopted but trust is collapsing. Usage is high, yet teams routinely doubt outputs and re-verify work, creating hidden operational friction. Without reliability and auditability, adoption rises while credibility falls, and trust becomes the real infrastructure of AI integration.davidnowak.meWhy Are Teams Using AI Tools They Fundamentally Don't Trust? - DAVID NOWAKTeams adopt AI at record rates but trust it at half that speed. Uncover the hidden 000
Bertrand Duperrin @duperrin.com · 11/12/2025Culture reshapes technology, not the reverse. Each innovation moves from fear to creative reuse, social filtering and slow integration. Even with AI, cultural resistance sets the pace, ensuring tools adapt to society rather than transform it overnight.gilescrouch.substack.comWhy Culture Eats Technology for BreakfastEvery technology since the Stone Age faces the same cultural gauntlet. None survive unchanged. 000
Bertrand Duperrin @duperrin.com · 11/12/2025AI adoption alone changes nothing. It produces scattered uses and local gains without reshaping work. Only true appropriation, which reorganizes activities and workflows collectively, creates lasting impact. Adoption is a surface metric; transformation comes from redesigned work.duperrin.comAI adoption does not replace productive appropriationBusinesses have always used the word "adoption" when talking about technology, especially when referring to what I call "freeform" technology. By this I mean to 000
Bertrand Duperrin @duperrin.com · 10/12/2025Microsoft place les jumeaux numériques au centre du futur du travail, en modélisant rôles, processus et connaissances pour déléguer des tâches à des agents. Au-delà de l’automatisation, c’est un nouveau paradigme collaboratif qui émerge. Reste la question clé des limites de cette délégation.fredcavazza.netDes intranets collaboratifs aux environnements numériques de travail augmentés par l’IA - FredCavazza.netL’évolution des environnements numériques de travail s’accélère avec l’émergence des « Frontier Firms« , ces organisations où collaborateurs et agents intelligents forment désormais des équipes hybrid... 000
Bertrand Duperrin @duperrin.com · 10/12/2025AI is widely adopted but barely productive. McKinsey shows 88% use it, yet only 7% fully deploy it. BCG sees a 5% elite capturing most gains. Both point to the same issue: firms aren’t structurally ready, and AI amplifies organizational weaknesses instead of fixing them.duperrin.comAdoption and impact of AI: lessons (and limitations) from the latest McKinsey and BCG studiesIn recent months, reports on the state of AI have been piling up, each heralding a major disruption in the way businesses work, make decisions, and create value 186
Bertrand Duperrin @duperrin.com · 09/12/2025IBM’s Arvind Krishna says the #AI spending race makes no economic sense. A 1-GW data center costs $80B; scaling to 100 GW means $8T, requiring $800B profit yearly. He sees #AGI as a near-zero-chance bet.futurism.comIBM CEO Says the Math Just Doesn't Add Up on Its Competitors' AI SpendingIBM CEO Arvid Krishna said that the math of its AI competitors' spending doesn't add up, and won't return a profit. 000