Ralph Van Coillie, MBA

McKinsey just released the “State of AI 2025” Report ↳ They reveal how businesses are adopting AI ↳ What separates AI winners from the rest ↳ The challenges businesses are facing TL;DR: → AI adoption surges (78% overall, 71% use generative AI). → Big companies lead, restructuring workflows and hiring AI talent. → CEO oversight boosts impact, especially in large firms. → Few track KPIs, slowing AI value capture. → AI reshapes jobs, driving reskilling over layoffs. → Revenue up in sales, supply chain; costs down in HR, IT. → Service jobs shrink, tech roles grow. → Success needs full AI transformation, not experiments. Key Takeaways: → AI adoption is rapidly growing, but most companies are still early in capturing value. → Large companies (> $500M revenue) are adopting AI faster and restructuring workflows. → CEO involvement in AI governance strongly correlates with financial impact. → Less than one-third of organizations follow best practices like tracking KPIs. → AI talent demand remains high, but hiring difficulty is easing slightly. → More employees are being reskilled rather than replaced. → AI is boosting revenue in marketing, sales, and supply chain. → Cost reductions are most visible in HR, risk & compliance, and IT. → Jobs in customer service and supply chain are shrinking, while IT and software engineering roles are growing. → Companies that fully transform their business with AI (rather than experimenting) will win. AI isn’t just a tool. It’s a business transformation strategy. Is your business adapting fast enough?

265

125 Comments

🚨 A 27 million parameter model just outperformed Claude 3.5 and Gemini on hard reasoning tasks. No chain-of-thought. No massive pretraining. No hallucinations. Meet HRM – Hierarchical Reasoning Model, a brain-inspired AI that might be our first real glimpse at AGI. Built by Sapient Intelligence (a Singapore startup founded by a Gen-Z Tsinghua prodigy + ex-DeepMind researchers), HRM doesn’t use tokens like traditional LLMs. It doesn’t “think out loud” by predicting the next word. Instead, it thinks internally. Like a human brain. - One module makes quick decisions - Another refines strategies over time And they loop—just like real cognition. Results? – 40.3% on ARC-AGI-1 (vs Claude’s 21.2%) – 100% on Extreme Sudoku – 100% on Maze-Hard All with just 1000 training examples and zero pretraining. This is called Chain-in-Representation which is a shift from CoT (Chain-of-Thought). No prompt hacks. No brute force. Just smart, recursive internal reasoning. Why it matters: 🔹 Structure, not size, might be the key to AGI. 🔹 HRM’s architecture spontaneously mirrors brain patterns seen in neuroscience. 🔹 It adapts—taking longer on complex tasks, faster on simple ones. In an era obsessed with trillion-parameter scaling and GPU burn, HRM is a wake-up call. A 27M parameter model… …trained on 1000 examples… …just beat some of the most expensive models ever built. OpenAI, Google, Anthropic: are you watching? We may have just seen the first crack in the Transformer empire.

99

18 Comments

From Data to Decisions — How Metadata Research Became Future-Ready AI Powering Decisions... Back in 2016–2019, during my PhD, I dived deep into something many overlooked: Metadata. While most were obsessed with just the text, I experimented with metadata — time, source, location, follower counts, retweets — training machine learning (ML) models to understand the true importance of keywords. Fast forward to today: AI has exploded, but guess what’s still powering the smartest systems? Not just the raw data — it’s the metadata that gives AI context, relevance, and decision-making power. Recent advancements have seen the rise of decision intelligence, a discipline that combines data science, social science, and managerial science to enhance decision-making processes. From trend detection to decision intelligence, metadata-driven models are shaping the future. The algorithm I built and trained back then wasn’t just academic — it was futuristic groundwork for today’s deep tech intelligence. By integrating AI and machine learning with decision intelligence, organizations can transition from data-driven to decision-centric models, ensuring that insights lead to actionable outcomes. As we move into an era of graph neural networks, temporal models, and metadata-enriched LLMs, the potential to turn raw streams into sharp decisions is bigger than ever. I bet on metadata and I am so glad that today, it’s driving the AI breakthroughs reshaping decision intelligence. ✨ The takeaway? Sometimes, the breakthroughs of tomorrow are hidden in the details we choose to explore today. Would love to hear — how are you using data, metadata and deep technologies to amplify your AI or business models? Let’s exchange ideas! 💬 #AI #DeepTech #Metadata #DecisionIntelligence #MachineLearning #Innovation #FutureOfAI #PhDResearch #AITrends #DataToDecisions #TechLeadership Dr.Panneerselvam ( PS ) Madanagopal Sanjeev K. Nayak (Ph.D., MBA) Tolga Turker Vikas Sood Purnima Jagannathan Meghali Sachdeva Dr. Seema Verma Apurva Durugkar

30

2 Comments

From Author to Lifelong Learner: My Data Integration Journey ... Two years ago, I had the incredible privilege of publishing "The Definitive Guide to Data Integration" with Packt, alongside my talented co-authors Pierre-Yves BONNEFOY, Emeric Chaize, and Mehdi TAZI Writing that book was a deep dive into helping organizations unlock the power of data integration to efficiently manage, transform, and analyze their data. It was both humbling and exciting to distill our collective knowledge into something that could help data professionals navigate this complex landscape. But here's what I've learned since then: Publishing a book doesn't make you an expert who knows it all — it makes you acutely aware of how fast our field evolves and how much there is still to learn. That's why I remain passionate about continuous learning in the data space. I'm constantly on the lookout for resources that keep me sharp, inspired, and ahead of the curve. Two resources that have become indispensable for me are Packt's dataPro and BI Pro newsletters. Each edition feels like a quick masterclass packed with: ✨ Actionable insights you can apply immediately 🔧 Emerging tools that are shaping our industry 💡 Expert techniques from practitioners in the trenches Whether you're working on data pipelines, exploring modern data stacks, or navigating the BI landscape, these newsletters make it easier to stay informed without the overwhelm. If you're looking to sharpen your skills, explore new ideas, and stay inspired in the ever-evolving world of data and business intelligence, I'd definitely recommend subscribing: https://lnkd.in/er7bydVM The best investment we can make is in our own continuous growth. What resources have been game-changers for you? #DataIntegration #DataEngineering #BusinessIntelligence #ContinuousLearning #DataProfessionals

25

5 Comments