Erdem Erkul, Founder and Chairman of Cerebrum Tech.
Today’s challenge isn’t accessing information, but taking ownership and acting on it. AI can complete tasks, but humans still need to direct and oversee these systems and, ultimately, take responsibility for the results. That’s where agency comes in.
“High agency” is a popular term in tech. But what does it actually mean? The word “agent” comes from the Latin “agere,” which means “to do.” In other words, human beings who act have agency.
According to George Mack, it is a skill that can be learned from small practices. People with high agency may start with a small step, but they do something. They don’t just wait. They take responsibility. They think clearly, are ready to act and do not give up at the first “no.”
In the digital era, high agency matters even more. Why? Because AI agents are becoming increasingly common. These agents can take actions and finish tasks. Sometimes they work with people, sometimes they ask questions before they finish a task and sometimes they work alone. Is taking initiative more important than intelligence now?
According to AI researcher Andrej Karpathy, agency is more powerful than intelligence. It is also harder to find. High-agency people are the “I’ll figure it out” type. They keep going when problems arise. They verify results, exercise judgment and take ownership of outcomes. While AI can do the work, much of the responsibility for decision-making still belongs to humans.
Why Does AI Make Agency More Important?
AI helps us do more work in less time. But is more work always better? A recent MIT study looked at this question by carrying out an experiment with more than 2,200 people who were asked to create several online ads. Some people worked with other people, while others worked using AI. The people who used AI made 50% more ads, and their ad scripts were also better. But the people who worked together created better pictures. Those who used AI altered their texts less, and their ads looked more alike.
AI can help us work faster, but we must not rely on it too heavily. To create truly unique results, we should start with our own ideas and use AI as a source of inspiration. AI can do a lot, but people must retain ownership of important decisions and take responsibility for the outcomes.
What Should Leaders Get Better At?
AI tools are becoming more powerful, making human contribution increasingly important. While these tools can help us get the work done, responsibility for the outcomes remains with us.
I’ve found that this is one of the most important qualities that drives progress. During times when resources are limited and uncertainty is high, taking responsibility, bringing the right people together and taking that first concrete step can make a world of difference.
When I encounter a problem, the question I ask myself is: “How can I solve this problem as soon as possible?”
Sometimes it’s a phone call, sometimes research, sometimes a small experiment and sometimes simply saying, “I don’t know this, and I need to learn it,” which marks the beginning of the solution. Then comes following up, getting feedback and adjusting the approach when necessary.
This is also why human skills are so essential. Leaders need to ask good questions, listen to different views and help people work together. AI can give us many answers, but it cannot decide what is best for every company or every person. Leaders must understand the needs of their teams, see problems early and make clear choices. They must also create an environment where people feel free to ask questions and share new ideas. This way, AI can support coming up with better solutions, while people keep the power to lead and take responsibility for the results.
This is why I do not see better AI tools as the real advantage. Every company can buy the same tools. The real advantage lies in the people who use these tools, recognize problems, take ownership and turn what is possible into measurable results.
This raises a question for every board and CEO: As AI gets better at execution, what should leaders get better at? I believe the answer has three parts. First, leaders must focus on the right problems. Investing in technology is not enough. Second, they must question how reliable AI outputs are, especially when the stakes are high. Third, they must protect accountability. A decision made with AI is still a human decision, and someone must answer for it.
How Can We Achieve High Agency?
Below are four simple steps to develop a higher level of agency. They are most useful when resources are limited and the future is not clear.
1. Define the problem. Ask yourself: “What can I do here?” Even in a hard situation, there is usually one thing you can control.
2. Take the first step. Do not wait for perfect conditions. It can be a phone call, an internet search or a small test. Choose a step that is small and easy.
3. Follow ideas. Many good ideas die after the first meeting, so follow up. Check the results, ask for feedback and change your method if it does not work.
4. Own the results. Good or bad, the result is part of your work. Take responsibility for both success and failure, and learn from the results.
As AI agents get better at doing tasks, human agency will matter even more. Our value will not be in doing every task ourselves; it will be in knowing what matters, making the hard calls and standing behind the results. AI gives us more power. High agency decides how we use it.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?