fastcompany10d ago
Below, Melissa Reeve shares five key insights from her new book, Hyperadaptive: Rewiring the Enterprise to Become AI -Native . Melissa was the first VP of marketing at Scaled Agile and thought leader in the SAFe in Marketing space. She went on to co-found the Agile Marketing Alliance. What’s the big idea? Most organizations are trying to bolt AI onto a system that was built for predictability. And it isn’t working. Pilots stall. Adoption plateaus. The organization gets faster at the edges, while the middle stays exactly as slow as before. What separates companies that succeed from ones that don’t isn’t the technology they choose, but rather the organization they become. Melissa calls those companies hyperadaptive . They’re architected to sense faster, learn continuously, and make smarter choices than any human could make alone. Listen to the audio version of this Book Bite—read by Melissa herself—in the Next Big Idea App , or buy the book . 1. Your operating system was built for the last century, and it can’t run AI. You can’t expect 21st-century results with an operating system built for the 20th century. However, there is a blueprint for getting from where you are to where you need to be. Let me explain what I mean by operating system . Most companies are still running on operating models built for the industrial era. Strategy flows top-down through layers of approval. Work moves sideways through functional silos. Hierarchy slows decisions. Handoffs lose information. This was the correct design for a world that valued consistency over speed. AI literally changes things. An organization that waits six weeks for a decision cannot compete with one that makes the same decision in six hours, with better data. Most leaders default to adding an “AI initiative” on top of the existing structure. With this approach, you end up with what Ethan Mollick calls the jagged edge : Some teams moving fast, while others remain stuck. Think about the companies that didn’t survive the digital transformation: Blockbuster , Kodak , Nokia . None of them died because the technology wasn’t available. They died because inertia kept the organization in place. With digital transformation, companies had about a 10-year window to figure things out. With AI, that window is closer to 18 months. So, how do you get from the operating model of today to an AI-native way of working? Hyperadaptive provides a five-stage path. The model is research-backed, specific, and already being used by leading companies. The companies winning with artificial intelligence have replaced the operating system underneath them, including the way the people, processes, and culture move together. There is a way to make these changes incrementally. You can start from where you are and bring the organization along, piece by piece. 2. AI doesn’t install itself. In the 1990s, when personal computers showed up at work, we didn’t put a PC on everyone’s desk and say, “Go have fun.” We trained people. We changed processes. We rebuilt how work was done. With AI, somehow, we’re trying to skip these steps. AI is like a piano. Anyone can walk up and start pounding the keys. That’s easy. But playing an actual song takes deliberate practice and guidance. AI is deceptively simple. The interface invites you in. However, the result you get without effort is mediocre. The result you get with the right structure and support can be transformational. “AI is deceptively simple.” Brad Miller was Moderna’s chief information officer during its AI transformation, and he said something that stuck with me. “90 percent of companies want to do generative AI,” he told me. “Only 10 percent succeed. The reason isn’t the technology. They haven’t built the mechanisms to transform their workforce.” That 10-to-90 gap is one of the most important numbers in this conversation. Moderna is in the 10 percent. In early 2023, its CEO, Stéphane Bancel, stood before his executive team and proposed something that sounded impossible: Bring 15 new drugs to market in five years. A single drug typically takes 10 years to develop and costs upward of 2 billion dollars. Bancel wasn’t asking his people to work harder. He was asking them to work differently, with AI as a coworker, strategic adviser, and accelerant. They stopped asking, “How does AI fit into our current way of working?” and started asking, “What’s the best way to work in an AI-powered world?” Six months in, Moderna had reached 100 percent generative AI adoption across the organization. They did that by building the mechanisms. Training. Coaching. Process redesign. A culture that treated AI fluency as a core capability, not an optional skill. If you want AI to transform your organization, you have to invest in the same level of ongoing training, coaching, and time to practice you’d invest in for any other major capability. 3. Learning is a bidirectional flywheel, not a curriculum. AI doesn’t stand still. The model your team trained on six months ago has been replaced twice. The prompts that worked in January won’t work in April. The use cases that were impossible to imagine last year are now table stakes. You cannot build a static curriculum for a moving target. So, forget the corporate training catalog. What you need is a learning arena, a place where people experiment, share, and build on each other’s experiments in real time. PwC figured this out. They run something called prompting parties. Yes, parties. Cross-functional groups come together, work through real business problems with AI, and walk out having taught each other things their training department couldn’t have built a course around. The learning is social, specific to the work, and spreads faster than any LMS could carry it. “The model your team trained on six months ago has been replaced twice.” But peer learning on its own isn’t enough. You also need a mechanism to capture what people are learning and feed it back into the system. This is what I call a bidirectional AI learning flywheel. AI Activation Hubs are small cross-functional pods that operationalize AI within a function, run experiments, and capture what works. AI Leads, who are your internal champions and automation translators, carry that learning to the front lines so people can apply it tomorrow. And critically, the front lines push their own discoveries back up to the hubs, where they get refined, tested, and pushed out across the rest of the organization. Learning, traveling in both directions, and compounding. Because AI itself is updating, the flywheel doesn’t only spread knowledge. It refreshes the knowledge as it goes. Organizations that create AI-powered learning loops to sense and respond in real time will lead the next decade. They are the ones who have built the infrastructure for people and AI to update each other faster than technology can change. If your AI training plan looks like a course catalog, you’re already lost. Build learning arenas. Build the AI flywheel. Make learning a system, not a syllabus. 4. Move one dimension and you get random acts of AI. Most AI initiatives are focused on tools. Pick the right model. Roll it out. Train people. Done. The problem is that an organization is a system. When you change one part of a system without changing the others, you get isolated successes—what I call random acts of AI. Pilots that don’t scale. Teams that get faster while other teams stay stuck. Productivity gains that disappear the moment people try to coordinate across functions. I spent a lot of years working in the transformation space. The Toyota Production System. Agile. DevOps. Every single one of them taught the same lesson. Progress stalls when you fail to move multiple dimensions in concert. For AI, the book lays out nine dimensions you must move together. Here are three that almost nobody is touching: Incentives. If your reward systems still pay people for being right rather than for learning fast, you will not become hyperadaptive. AI work involves unknowns. People have to feel safe to try things that don’t work. Decision rights. AI collapses decision hierarchies. A junior analyst with the right model can now make a call that used to require three layers of approval. If you haven’t rewired who decides what, you leave a lot of speed on the table. How you organize. Functions versus value streams. Permanent teams versus dynamic ones. Most organizations were built around work as it existed 20, even 40, years ago. AI requires you organizing around the work as it exists now. Organizations tend to move slowly and unevenly. The five-stage road map accounts for this. At each stage, you move the dimensions that are ready to move. They don’t have to move in lockstep, but the dimensions do have to be considered as a system. Let one dimension get too far behind, and it blocks progress in the other dimensions. Treat AI as a tool initiative, and you get tool results. Treat AI as a system to be reinvented, and you get organizational results. 5. History tells us where the jobs go, but who’s responsible for getting people there? The World Economic Forum’s “Future of Jobs Report” projects that 92 million jobs will be displaced by 2030. Jobs disappearing is what makes the headlines. And that number deserves to be taken seriously. What doesn’t make the headlines is that the same World Economic Forum projects that 170 million new jobs will be created in that same window. Net positive 78 million. The question isn’t whether work is going away. The harder question is where it’s going, and whether we’re paying attention. History tells us where it goes. Electricity. Factory automation. DevOps. The introduction of personal computers in the workplace. Each of these revolutions followed the same pattern. People stopped doing the task by hand and began building, monitoring, and maintaining the systems that performed it. The jobs evolved. Some industries were hurting for a long time. The macro picture, every single time, was net positive growth. “The harder question is where it’s going, and whether we’re paying attention.” Who is responsible for getting people across that bridge? The government? Individuals? Companies? Smart companies have already made that choice. They calculated the cost of firing one workforce and hiring another—not just the recruiting expense, which is significant, but also the institutional knowledge they’d lose, the customer relationships, and the cultural memory. Leading companies like Unilever recognize the cost of this displacement and are investing in upskilling and AI matching. They use AI to identify which existing employees can be reskilled for which emerging roles and make the investment. They’re treating it as strategy, the same way they’d treat any other long-term investment. The pattern of where jobs go is clear. The data is on our side. And the companies that are choosing to take responsibility for their people are doing it for the same reason they make any other long-term bet: Because it pays off. AI is going to reshape the work. What’s up to you is whether you become the company that helps your people make that jump, or the company that loses them and then has to find them again after your reputation has taken a hit. This article originally appeared in Next Big Idea Club magazine and is reprinted with permission. Enjoy our full library of Book Bites—read by the authors!—in the Next Big Idea app .
