The Skill of the Future: “Agency”
One of the most frequently asked questions lately seems to be this: “Which skills will matter most in the future?” The intention behind the question is right—but the question itself is somewhat incomplete. Because in an agentic world, what truly makes the difference is not which skills you possess, but what you are able to do with them. More precisely, it is how much agency you have. By “agency,” I mean the capacity to recognize problems, take ownership, generate solutions, and turn those solutions not into one-off actions but into sustainable and scalable systems.
One of the core lessons Agile taught us was that value emerges not from silos but from flow. That’s why cross-functional teams emerged—bringing business, technology, design, and operations to the same table. Instead of waiting for approvals, teams took ownership, and feedback and learning accelerated. Yet today we are facing an important inflection point: these structures still operate largely at a human scale. The agentic world arrives with a scale that is faster, more parallel, and more continuous. For this reason, agentic teams should be seen as the natural next evolution of agile and multidisciplinary ways of working.
With this evolution, the way we solve problems is also changing. In the past, humans solved problems; today, humans and agents solve them together. Roles are becoming clearer: humans remain the owners of decisions, while agents take responsibility for speed, scanning, and computation. This distinction is critical because agency is being redefined precisely here. Agency no longer means doing everything yourself; it means ensuring that the right work is done by the right actor at the right time. That’s why agentic thinking is not a break from Agile—it is Agile at a new scale.
So what does “skill” mean in an agentic world? It is not simply knowing a tool—nor is it just writing prompts. Saying “we use AI” is already a baseline statement. The real differentiator is having the natural reflex to ask: “How do we solve this problem as a system where humans and agents collaborate?” High-agency structures define problems from the right place, maintain cross-functional connections, eliminate repetitive work from the system, and continuously improve flow. What we are talking about here is not a checklist of competencies but an organizational reflex.
We can still talk about a few core capabilities that build this reflex and cover the end-to-end cycle of agency. The first is clear problem definition. Agency always begins with the problem. A poorly defined problem will lead you in the wrong direction, no matter how fast you move. In the agentic world, this becomes even more critical—because agents will scale the wrong solutions much faster. That’s why starting with the user’s pain, rather than the solution, is vital.
Another key capability is flow design. In the past, we designed human-to-human workflows; today, we design human–agent flows. At which points do humans step in? Which decisions remain human? Which scanning, analysis, or recommendations are delegated to agents? The goal here is not to accelerate existing work with AI but to redesign the work itself based on AI’s new capabilities.
The third capability is human–agent orchestration. Think about how team members synchronize within a Scrum team; in the agentic world, a similar synchronization must exist between humans and agents. This is less a technology issue and more a matter of leadership, culture, and design. Working with a few agents today may feel manageable—but when organizations begin working with hundreds of agents, orchestration will become a defining capability.
Finally, there is the capacity to build systems. Agile approaches learning through continuous observe-and-adapt cycles. The agentic approach extends this learning beyond individual team practices and embeds it into the organization’s decision-making and problem-solving architecture. Building systems here does not mean adding more processes or imposing solutions; it means clarifying which shared principles and frameworks guide recurring problems and decisions. Agency becomes tangible when learning does not get lost, context is preserved, and adaptation relies not on individual effort but on a shared way of thinking.
Defining agency only as taking action would be incomplete. Agency also includes the capacity to keep experimenting without giving up. When something doesn’t work, pausing to reflect, making small adjustments, and trying again is a natural part of the process. Learning and doing do not separate here—they move together. In agentic structures, the real risk is not making mistakes, but returning to familiar yet inefficient patterns under the excuse of “it’s not working.” Agency grows through the courage to keep learning within uncertainty.
Today, we have access to almost any knowledge we might need—how-to videos, guides, models… Yet as Dan Koe also points out, many individuals and organizations do nothing with that knowledge. This is a critical insight. The path to success has never been more accessible technically—yet the number of those producing real outcomes is not increasing at the same pace. The reason was never just access. The real differentiator has always been agency. High-agency structures act without waiting for permission and treat obstacles not as excuses but as design inputs. If resources are limited, goals may shrink—but direction does not get lost. Intermediate milestones are consciously chosen to enable the next step. That’s why agentic organizations move much faster than others, even when they have access to the same information. Because the real issue is not what you know, but what you activate with what you know.
Agile and cross-functional ways of working brought us to the right place. But stopping there is not enough. The winners will not be those who simply add more skills or use more tools. The winners will be those who can design humans, agents, and flow architecture together. The future does not need more skills—it needs far more agency.
Comment(s)