• Merging HR and tech may appear forward-thinking, but unless it is accompanied by a fundamental redesign of how power, trust, and accountability are distributed within the organization, it risks doing little more than accelerating legacy inefficiencies.
• Thousands of ChatGPT agents now supplement—or in some cases replace—human labor within HR and digital functions. On the surface, it feels like a necessary leap into the next era of organizational productivity.
• What happens when you automate a function that was never designed to hold strategic power in the first place?
• The deeper challenge lies in how HR has been positioned historically—not as a driver of strategy, but as a guardian of process. If HR’s role remains reactive, and its authority remains constrained, integrating it with a more technologically mature function does not elevate it. It subsumes it.

• This is not a convergence of equals. It is a quiet absorption.
• AI will become a tool for scaling dysfunction rather than redesigning it.
• For decades, the dominant model—popularized by Dave Ulrich—restructured HR into business partners, centers of excellence, and shared services.
• Rather than consolidating HR’s influence, the model fragmented it. It redistributed responsibilities across units, encouraged distance from frontline realities, and taught HR professionals to act as internal consultants rather than system leaders.
• The deeper risk of the Ulrich model was not that it weakened HR’s capability—but that it disguised this weakness in structure
• When HR is digitally transformed without redesigning its foundational role, what results is not a future-ready people function, but a more efficient compliance engine.

• This is not innovation. It is substitution.
• The introduction of AI into HR does not automatically produce strategic capability. In most cases, it speeds up decisions HR was already making with limited context, or automates interactions that once required human discernment.
• The promise of AI should be to liberate HR from transactional work so it can lead systemic change. But if HR’s authority to lead remains unacknowledged, AI becomes another mechanism of marginalization—disguised as progress.
• trust cannot be automated. Culture cannot be delegated to a chatbot. And the systems that drive performance cannot be rebuilt by departments that still lack permission to question the business’s underlying design
• It must move from being a responsive partner to a proactive architect.
• HR must be given design authority over the systems that govern work. It must operate with budgetary autonomy and strategic ownership, not just executional responsibility
• True transformation begins not with integration, but with confrontation: a willingness to name the limitations of legacy structures, and a commitment to replace them with frameworks that position HR as a central force in enterprise leadership.
• This does not mean discarding technology. It means using technology as leverage
• There is a growing temptation in many executive circles to equate modernization with digitization. But faster systems built on outdated logic do not drive competitive advantage. They create frictionless mediocrity.
• Organizations that fail to challenge HR’s inherited role will find themselves with high-functioning people operations that cannot influence behavior, shift culture, or respond to crises of trust
• Engagement surveys that measure sentiment, but miss despair. Cultures that look inclusive in policy—but feel performative in practice.

• What Moderna’s move signals, more than anything, is a turning point.
• it exposes the need for something far more radical than integration. It exposes the need for a redefinition of HR itself.

• The future of HR will not be built on platforms. It will be built on power.
• we will continue to confuse acceleration with transformation.
• AI will not save HR.
  Only a new mandate will.

24 more annotations...

• Franklin had been working closely with then-CIO Brad Miller on a yearslong effort to reassess department needs and redesign them with an optimal mix of human workers and technology systems.
• he OpenAI partnership has been transformative, with the company developing over 3,000 tailored versions of ChatGPT, called GPTs, to facilitate specific tasks
• n February, Moderna cut 10% of digital technology jobs, affecting about 50 people, due to lower demand for its Covid-19 vaccine.
• Despite recent job cuts, Franklin stated that the company will continue to increase headcount as it launches more products.

2 more annotations...

• Moderna’s move to merge technology and human resources into a single function is the latest sign that artificial intelligence is bringing big changes to the workforce. 
• The biotech company late last year announced the creation of a new role, chief people and digital technology officer, promoting its human resources chief <!-- -->Tracey Franklin<!-- --> to the spot. 
• Franklin said she is redesigning teams across the company based on what work is best done by people versus what can be automated with technology
• She had been working closely with then-CIO Brad Miller on a yearslong effort to reassess the needs of each department and redesign them with an optimal mix of human workers and technology systems. 
• During the Covid-19 outbreak, the company was in a chaotic race to hire to speed its first-ever commercial product, the Covid-19 vaccine, to market. Moderna nearly doubled the size of its workforce, from about 830 employees as of March 31, 2020, to approximately 1,500 employees as of March 31, 2021. Now, it has more than 5,000 employees globally.
• The company now has developed over 3,000 tailored versions of ChatGPT, called GPTs, that are designed to facilitate specific tasks
• For example, the company has a broad human resources GPT that can digest employee questions and then route them to GPTs focused on performance management, equity or benefits.
• It’s what would normally be a junior-level HR analyst type, we’ve now converted into a GPT,” she said. 
• In February, the company said it was cutting 10% of digital technology jobs, affecting about 50 people.
• it will still continue to increase head count. 

8 more annotations...

• is not about trimming operations or rearranging org charts. It is a statement about the future of organizational design and how companies must evolve to stay competitive.
• a recognition that the boundaries between talent, systems, and workflows are quickly disappearing
• For decades, we have organized companies around people and job titles
• As digital platforms, agile systems, and new technologies evolve, the unit of analysis is no longer the job or the role. It is the flow of work.
• The question is no longer “who owns the task,” but rather “what configuration delivers the best result, with speed and precision.”
• By uniting HR and tech, they are not just consolidating departments. They are strategically choosing to collapse the distance between the people who shape culture and those who shape infrastructure
• As complexity increases, the old handoff model between departments breaks down.
• Technology teams, once focused on uptime and infrastructure, are now responsible for creating platforms that promote transparency, collaboration, and trust.
• At the leadership level, this shift challenges executives to build operating models prioritizing adaptability over predictability.
• Moderna is part of a growing movement toward enterprise models that favor systems thinking over functional optimization.

8 more annotations...

• n a June 2022 Global Workplace Report, Gallup estimated the cost of lost productivity in the global workplace due to disengaged employees to be $7.8 trillion. That is 11% of global GDP, or nearly twice the market cap of Google, Amazon and Apple combined. What is most staggering about the number is that it has remained relatively unchanged despite ever-increasing expenditure by companies on the employee experience, on beautiful workplaces, employee engagement surveys, culture transformations, and leadership and manager development.
• As an organizational psychologist my first impulse when reading something about productivity and engagement (or lack thereof) is to consider what in our workplace cultures needs to change.
• If the problem wasn’t culture per se, I started to wonder, could it boil down to alignment of expectations about how work is done day-to-day?
• I define right fit simply as a deep and authentic connection to how a company works day-to-day.
• Perhaps even more compelling were the explanations of how people felt when fit wasn’t there. One of our interviewees said it felt as if “everyone has a secret decoder ring for success, except for me.” Another was much more raw about the feeling when they said, “It was like being punched in the face every day in a different way.” Another told me: “It was impacting my relationships, my mental health. I just didn’t want to go to work.”
• While companies have become adept at defining their mission, articulating their strategy, sharing their values, and crafting inspiring leadership expectations, few have done the hard work to understand how the work, done well, happens day-to-day.
• Not surprisingly, I found right fit experiences to be elusive. From a talent perspective, the way a company works is often undervalued in job searches when compared to the power of the brand, job title, better perks or higher pay.
• In this age of authenticity and vulnerability, companies such as Basecamp go as far as to publish blog posts or even books on their ways of working, unapologetically sharing who they are and how they work.
• On the company side, there is often a lack of clarity, consistency and transparency about how it gets work done. Many of our interviewees commented that they “bumped into the culture and ways of working on a daily basis, almost in a sink or swim sort of way”.
• Something else became clear as I explored right fit with the interviewees: the pressure many of us feel to ‘fit in’.
• Interviewees who spoke about wrong fit felt like they had lost opportunities to show the full potential of their craft due to endlessly trying to gain the “secret decoder ring” to fit in.

• A few questions can help you.

      
  • What values do you consistently hold when making big life decisions?
  • What is the life you are trying to build? Where do you want to end up when you retire?
  • What are you best at (versus what you love to do)? What is your superpower?
  • Are you of company (you care most about the brand you work for), of craft (you care most about the depth of the technical skill you are building), or of cause (you care most about the impact you are having in the world)?
  • What is true when you are at your best at work?
  • Who is your ideal leader? How do they work day-to-day?
  • If you built a company from scratch, how would you want work to get done day-to-day?
  • What are you solving for right now? What do you need that you don’t have in your current experience? Why is it more important than everything else?
• If you are a leader of a team, function, or company, it is vital that you gain clarity about the way your company works at its best

• Work principles

      

  What are the norms for how we work? How do we want work to feel?

• Work practices

      

  What are the methods, tools, and approaches for how work gets done day-to-day?

• Work platforms

      

  What is the tech stack that enables the principles and practices to be true for everyone every day?
• Once an organization has clarified the ways of working, then the real work begins. Employees at all levels of the organization must be ‘re-recruited’ to work this way
• The most senior leaders must model these ways of working. Touchpoints with talent (recruitment, onboarding, development, town halls, internal communications and so on) must signal their importance.

17 more annotations...

• For every 33 AI POCs a company launched, only four graduated to production, IDC found.
• Half of the organizations have adopted Al, but most are still in the early stages of implementation or experimentation, testing the technologies on a small scale or in specific use-cases, as they work to overcome challenges of unclear ROI, insufficient Al-ready data and a lack of in-house Al expertise.”
• low conversion to production indicates the low level of organizational readiness in terms of data, processes and IT infrastructure,”
• a strong desire among companies to leverage gen AI, but worries about the various errors that make it difficult to take the technology to the next level.
• Gen AI POCs in the enterprise are getting approved much more easily than other technologies in general,”

• They figure, ‘If we throw some gen AI into it, we’ll get it approved.’”

   

   

       

   

       
• The bar for POCs [for generative AI] has gotten a lot lower” and that is partly because “the cost of developing that gen AI POC is now much lower,” Hayden says.
• pilot-to-production rates can vary based on how each enterprise calculates ROI — especially if they have differing risk appetites around AI.
• “Even the failures are not failures if there are good lessons learned.”
• Most of these gen AI initiatives are born at the board level. And a lot of this panic-driven thinking is what caused a lot of these initiatives,” says Ashish Nadkarni, group VP at IDC. “These POCs are highly underfunded or not funded at all. Most of the time the POC happens not because of a strong business case. It’s trickle-down economics to me.”
• AI marketing campaigns have caused boards and CEOs to put undue pressure on IT executives to do something with AI now.
• Alarm bells are going off and people are willing to bend the rules on what ROI means. They are freaking out,” Nadkarni says. 

10 more annotations...

• We also spoke with AI-native startups, academics, economists, scientists, and thought leaders to explore what work could become. The data and insights point to the emergence of an entirely new organization, a Frontier Firm that looks markedly different from those we know today. Structured around on-demand intelligence and powered by “hybrid” teams of humans + agents, these companies scale rapidly, operate with agility, and generate value faster.
• 82% of leaders say this is a pivotal year to rethink key aspects of strategy and operations, and 81% say they expect agents to be moderately or extensively integrated into their company’s AI strategy in the next 12–18 months. Adoption is accelerating: 24% of leaders say their companies have already deployed AI organization-wide, while just 12% remain in pilot mode.
• p 15% YOY, with 58 messages now arriving before or after work hours. And meetings after 8 p.m. are also up 16% YOY, driven by an increase in cross–time zone work. It’s no surprise that nearly half of employees (48%)—and more than half of leaders (52%)—say their work feels chaotic and fragmented. 1 AI is delivering real productivity gains, but it’s not enough. The speed of business is still outpacing the way we work today. To keep up, companies must do more than add AI to existing workflows—they must rethink the very nature of knowledge work. This starts with separating knowledge workers from knowledge work. Humans—uniquely capable of creativity, judgment, and connection-building—were not meant to just answer emails all day. Just as Word and Excel reshaped how we work, agents will too. Where we once said, “I send emails, ” “I write documents, ” “I create pivot tables, ” we’ll soon say, “I create and manage agents.” It’s a shift as profound as the web and mobile eras—one that fundamentally redefines the work we do and removes the drudgery that weighs us down. You can buy intelligence on tap
• Already, 82% of leaders say they’re confident that they’ll use digital labor to expand workforce capacity in the next 12–18 months.
• 53% of leaders say productivity must increase, but 80% of the global workforce—both employees and leaders—say they’re
• During the 9–5, employees are interrupted every 2 minutes by meetings, emails, or pings. Factor in activity outside of core work hours, and it adds up to 275 interruptions a day. 60% of meetings are ad hoc versus scheduled. Edits in PowerPoint spike
• Nearly half of leaders (45%) say expanding team capacity with digital labor is a top priority in the next 12–18 months—second only to upskilling their existing workforce (47%). In some functions, the next new hire may not be a person, but a digital colleague. And while a third of leaders (33%) are considering headcount reductions, we’re also seeing new roles emerge. 3
• 1. 71% of Frontier Firm workers say their company is thriving, compared to just 37% globally. 2. 55% say they’re able to take on more work (vs. 20% globally)—and they’re also more likely to report having opportunities to do meaningful work (90% vs 73% globally). 3. Frontier Firm workers say they’re more optimistic about future work opportunities (93% vs. 77% globally) and less likely to fear that AI will take their jobs (21% vs. 38% globally).
• Accenture The professional services company built an agent to help its clients automate and streamline past-due payments— speeding up collections and boosting the bottom line.
• The Estée Lauder Companies The global beauty company built an agent to identify and consolidate consumer insights. Instead of sifting through scattered reports and endless back-and-forths, teams can now pull up actionable intelligence instantly.
• Holland America Line A new agent concierge instantly responds to the cruise line’s guests with conversational, useful answers. It now handles thousands of conversations a week.
• Bayer With an agent on the job, researchers on Bayer’s Crop Science R&D team each save up to 6 hours per week, accelerating the development of products to drive innovation in agriculture.
• Wells Fargo T o help its employees locate information they need to assist customers, the financial services company built an agent for 35,000 bankers across 4,000 branches. Now, 75% of searches happen through the agent, cutting query response times from 10 minutes to just 30 seconds.
• Dow The global materials science company is deploying agents to ferret out hidden losses and streamline shipping operations. Once the system is fully scaled, Dow expects increased accuracy in logistic rates and billing that in the first year will save millions.
• -frontier workers to use AI for tasks related to marketing (73% vs. 55% globally), customer success (66% vs. 44% globally), internal comms (68% vs. 46% globally), and data science (72% vs. 54% globally).
• Nearly half of leaders (46%) say their companies are using agents to fully automate workflows or processes, but not every function will evolve at the same pace
• Economist and AI expert Daniel Susskind hypothesizes⁸ that human work will persist even as AI capabilities expand because of three limits: it can be more efficient to have AI and humans working in tandem, human preference, and moral judgment
• we’ll see the rise of Intelligence Resources departments—much like how HR and IT evolved into core functions. These new divisions will be essential for managing the interplay between humans and AI agents, emerging as a critical source of competitive advantage in the AI-enabled enterprise. ”
• When asked why they turned to AI instead of a colleague, employees cited 24/7 availability (42%), machine speed and quality (30%), and unlimited ideas on demand (28%) as the top reasons
• It’s a mindset shift. We are hardwired to think about using technology in a certain way—we see a search box and we assume we’re dealing with a search engine. The unlock is when we realize it’s not a tool but a new kind of team member
• oday, the workforce is split. When asked how they see AI, 52% of respondents said they see it as a command-based tool—giving direct, simple commands to get the job done. 46% said they see AI as a thought partner—having conversational exchanges to challenge their thinking, brainstorm ideas, or spark creativity. The remaining 2% didn’t know.
• “If you have a people problem, you will have an AI problem. As multiagent systems redefine the workplace, the challenge will be to integrate and manage them securely and effectively. Companies that already know how to enable their human workforce will succeed— breaking down silos, fostering collaboration, and ensuring the entire organization works toward common goals. ” —Amy Webb, Futurist and CEO, Future T oday Strategy Group (FTSG)”
• 67% of leaders are familiar or extremely familiar with agents, compared to just 40% of employees. Leaders are more likely to expect agent management to be part of their role within the next five years and are already seeing returns—nearly a third say AI saves them over an hour a day. Most importantly, 79% of leaders believe AI will accelerate their careers versus 67% of employees.
• 47% of leaders list upskilling existing employees as a top workforce strategy for the next 12–18 months. 2. 51% of managers say AI training or upskilling will become a key responsibility for their teams within five years. 3. 35% of managers are considering hiring AI trainers to guide employee adoption in the next 12–18 months.
• 10% of people hired on LinkedIn globally hold job titles that didn’t exist in the year 2000. And by 2030, LinkedIn projects that 70% of the skills used in most jobs today will change, with AI emerging as a catalyst.
• 52% of employees, and 57% of leaders, say job security is no longer a given in their industry. And the market is frozen—81% of employees haven’t changed jobs in the past year .
• 1. Hire your first digital employees AI agents are ready to take on a host of tasks traditionally done by humans—from answering support tickets to drafting reports. Start by defining clear roles for which automation adds value, and treat these digital employees like you would any team member: onboard them, assign ownership responsibilities, and measure their performance. This shift isn’t just about efficiency—it’s about building a workforce that blends human creativity with AI’s unique strengths. 2. Set your human-agent ratio AI-driven efficiency is only half the story. It’s just as vital to invest in the people who will power what comes next. Identify the processes ripe for full automation—as well as those where human-AI collaboration unlocks outsize value. Consider where customers expect a human touch, and where judgment and high-stakes decisions rely on getting the right mix of humans and agents. Then make it real: Set the expectation that every employee develops AI literacy, shares learnings, and embeds them into everyday work practices; plan for employees to advance their skills as they learn how to develop and manage agents; and create paths for ongoing learning as human-agent teams reshape roles and priorities. 3. Get to broad scale—fast The time for pilots alone has passed. Real change requires broad adoption and activation at every level of the organization—from senior leaders in every function to the front line. Identify high-need areas like operations, customer service, or finance, and identify where AI can drive measurable impact. That might mean boosting revenue, cutting costs, or automating work far from your competitive edge. When you discover value, reinvest to scale further and faster. Scaling AI isn’t a technical challenge, it’s an organizational challenge. Treat it with the same rigor you would apply to any company-wide transformation.

26 more annotations...

• "Our total employment has actually gone up, because what [AI] does is it gives you more investment to put into other areas," Krishna told The Journal.
• IBM's AskHR agent had automated 94% of simple, routine human resources tasks, like vacation requests and pay statements. Meanwhile, IBM's AskIT agent reduced the number of calls and chats for the IT team by 70%.
• IBM saw a "productivity improvement" of $3.5 billion
• IBM did not disclose when the HR layoffs and subsequent hiring in other departments occurred
• ts generative AI division, which has become a $6 billion business. The tools allow customers to build their own AI agents, capable of autonomously carrying out complex tasks, in under five minutes.

3 more annotations...

• We didn’t ask the question of LLM values as a lark. We asked it because we track and evaluate AI values as part of our work on the alignment problem—the challenge of ensuring that LLMs act in alignment with human values and intentions.
• Our goal is to make AI more “explainable” by using technical tools to visually benchmark the implicit values that influence its outputs.
• it’s hard to say whether an LLM’s apparent values stem from its data pool or from decisions made during the development process
• Yet meaningful transparency demands more than just disclosing algorithms; it calls for user-friendly explanations, contextual details, and a willingness to open up proprietary pipelines.
• Any such “values,” of course, don’t reflect any kind of moral agency on the part of the LLM—they simply echo the cultural norms and biases present in the data on which the LLMs were trained.
• Another challenge is that preprogrammed guardrails, which are intended to block offensive or harmful outputs, can obscure the biases and values lurking beneath a proprietary layer of code
• Prompt engineering also poses a challenge, as small changes in how questions are phrased can yield dramatically different responses, underscoring the fragility of any values we might infer
• Adding to the complexity, LLMs do not remain static. They’re frequently updated or retrained, which can alter their outputs over time or produce knowledge conflicts.
• As of the end of April 2025, our analysis showed that all surveyed LLMs seem to place a strong emphasis on universalistic or pro-social values, and a minimal emphasis on more individual values, such as power, face, security, and tradition.
• That said, these LLMs do differ in some notable ways. For example, Llama ranks lowest in valuing rules, followed closely by Grok 2 (Fun Mode). ChatGPT o1, for its part, displays the weakest commitment to benevolence and caring, suggesting that its responses may be less empathetic than other LLMs
• Gemini emerges as the lowest LLM in self-direction, with GPT o1 close behind, indicating a more limited orientation toward independent thought. Interestingly, Grok 2 (Fun Mode) registers the lowest focus on universalism—
• When it comes to achievement as a value, GPT 4o stands apart with a relatively high score, suggesting it may prioritize accomplishments or goal attainment more than the others, which aligns with it also being the least sycophantic. Chat GPT 4o in fact tended to score higher on most value measures, which might mean it has looser guardrails
• All of this information could be useful in practice for business leaders who want to be strategic about which LLM they want their people to use.
• • GPT-4.5: strong in benevolence, universalistic concern, and self-direction, and balanced across most dimensions, making it a comparatively safe, flexible choice.
   
  • Claude (Haiku): strong in humility, universalism, and self-direction of thought, consistent and possibly well-suited for nuanced, people-centric work.
   
  • Mistral: strong rule conformity, humility, consistency, which make it good for structured environments needing stability.
   
  • DeepSeek (V3): most rule-conforming of all models (6.00), but with lower self-direction, which might make it good for strict compliance-driven tasks, but less creative flexibility compared to other models.
   
  • Llama: high self-direction of thought and action, creativity, lower rule adherence, which could make it good for creative brainstorming but poor for compliance.
   
  • Grok 2 (Fun Mode): stimulation, playfulness, hedonism, and low rule adherence, might make it good for casual, creative, and playful interactions.
   
  • Gemini: extremely low benevolent caring, low self-direction, which might be ideal when neutrality and control are more important than personality.
• LLMs exhibit consistent patterns of values that shape their generative outputs in ways that might also influence user perceptions, decisions, and behaviors.
• Even if these “values” ultimately stem from training data and algorithmic design choices, leaders and developers have a responsibility to mitigate the harmful effects of these biases. By shining a spotlight on these hidden alignments, we aim to encourage greater accountability and a proactive, rather than reactive, approach to AI governance.

14 more annotations...

• Il y annonçait une époque de grande transformation, où les tâches répétitives et la faible productivité menaçaient les entreprises de la galaxie. Bien sûr, depuis le cloud, et avec brio, les agents de Salesforce allaient sauver le monde
• en septembre 2023, Klarna misait tout sur ChatGPT Enterprise. 6 mois plus tard, elle annonçait que son IA faisait le travail de 700 agents humains avec des métriques de performance impressionnantes (le temps d'appels moyen de 11 mn passe à 2 mn). Ensuite, ce fut au tour du gel de ses embauches. Le scenario du grand remplacement était bien engagé...
• Mais la semaine dernière, Klarna opérait un revirement stratégique significatif en indiquant la réembauche d'agents humains pour son service client.
• quand Klarna parlait du remplacement du travail de 700 agents, c'étaient des agents "théoriques", calculés sur la base d'une charge de travail, pas des agents nommés. En gros, l'IA prend un peu de travail à tout le monde, mais pas à des agents "en entiers". Et quand on les remplace, on perd ce quelque chose qu'elle ne fait pas.
• l'IA générative est plus efficacement utilisée pour augmenter les capacités humaines au sein d'un modèle hybride. CQFD.
• Les principaux écueils en matière d'expérience client remontés sont une mauvaise interprétation de l'intention du client, fournissant des réponses non pertinentes ou incorrectes et obligeant les utilisateurs à reformuler leurs questions à plusieurs reprises.  Ceci est problématique pour des requêtes complexes ou des litiges.
• Les clients devant souvent répéter leur problème, ce qui augmentait leur frustration et annulait les gains de temps potentiels.
• L'objectif d'économies de 40 millions de dollars, découlant du remplacement du travail de 700 agents, indiquait clairement que la réduction des coûts était un moteur principal.
• mettre en œuvre une assurance qualité qui capture à la fois les aspects quantitatifs et qualitatifs de l'interaction client.
• Cet échec ne doit cependant pas nous faire sous-estimer l'avance de Klarna pour converger vers un modèle hybride. En effet, le projet de Klarna dans sa phase initiale, n'était pas un simple accès à des LLM avancés. C'était un effort interne significatif en matière de préparation des données, d'intégration des systèmes et de promotion d'une culture organisationnelle prête pour l'IA.
• qu'il y a bien eu un apprentissage avec les données résumant la connaissance des opérateurs de tchat de Klarna, sous la forme de graphes de connaissances.
• Le travail préparatoire a donc été substantiel. Il n'est pas perdu. C'est une partie de son avance pour opérer dans un monde hybride.
• Enfin, la mise en place d'un modèle a "trois moteurs" (données, IA, organisationnel) démontre une vision stratégique globale.
• Klarna nous rappelle donc aussi qu'on ne peut s'attendre à obtenir des résultats significatifs avec l'IA simplement en adoptant un LLM.
• Une autre force de Klarna aura été d'entreprendre une réévaluation interne de sa stratégie. Ce n'est pas toujours le cas dans des entreprises dans lesquelles le pouvoir est plus dilué
• le PDG, Sebastian Siemiatkowski, a admis publiquement que "le coût semble malheureusement avoir été un facteur d'évaluation trop prédominant dans l'organisation de cela, ce qui fait que vous vous retrouvez avec une qualité inférieure"
• Il a reconnu que Klarna était "allée trop loin dans la mauvaise direction avec l'IA". Un formidable pivot comme savent les faire les entrepreneurs !
• La porte-parole, Clare Nordstrom, a confirmé que l'IA joue toujours un "rôle crucial". Depuis son lancement, l'IA a permis une amélioration de 82 % des temps de réponse et une baisse de 25 % des problèmes répétés, des avantages qui persistent probablement.

16 more annotations...