Rethinking Soft Skills in the Age of AI

As AI becomes more capable, the skills we have traditionally called “soft” may turn out to be some of the most important capabilities people bring to work. 

For years, organizations have used the term “soft skills” to describe capabilities such as communication, empathy, adaptability, critical thinking, collaboration and judgment. The phrase has always been slightly misleading, because there is very little that is soft about giving difficult feedback, managing competing expectations, leading people through uncertainty or making decisions when there is no perfect answer. 

These capabilities are not secondary to performance. They shape how work gets done, how decisions are made, how people experience leadership, and how effectively teams move from discussion to action. 

As artificial intelligence becomes a more normal part of work, it may be time to reconsider the way we think about these skills altogether. 

Much of the current conversation about AI understandably focuses on what the technology can do. Organizations want to know which tasks can be automated, which roles will change, how productivity may improve, and what new technical skills employees need to develop. 

Those questions matter, but they are only part of the conversation. Another important question is what becomes more important for people as technology becomes more capable. 

AI can already summarize information, analyze data, generate ideas, compare options, prepare drafts, and suggest solutions with remarkable speed. In many situations, it can help people arrive at a useful first answer faster than ever before. 

What it cannot remove is the need to decide whether that answer is good enough, appropriate for the situation or worth acting on. 

This is the point at which the skills we have traditionally called soft start to look very different. 

When answers become easier, judgment matters more 

One of the biggest changes AI brings to work is not simply that information is easier to access. It is that convincing answers are easier to produce. 

A report can look polished even when the reasoning behind it is incomplete. A recommendation can sound confident while being based on weak assumptions. A piece of communication can be professionally written while completely missing what the person receiving it actually needs to hear. 

This changes what is expected of the person using the technology. 

The question is no longer only whether we can produce an answer. We also need to know whether the answer makes sense, what may be missing, what assumptions sit underneath it, and whether it fits the reality of the situation in front of us. 

That is what critical thinking looks like in practice. 

It does not mean questioning everything or making straightforward work unnecessarily complicated. It means being able to look beyond the first answer and notice when something does not quite add up. It means understanding that two situations may look similar on paper and still require different responses. It means recognizing that data can be accurate while the interpretation of that data may still be incomplete. 

Sometimes good judgment means knowing that an answer is good enough and moving forward. At other times, it means slowing down, asking another question and looking more closely. 

As AI gives us more answers, knowing when to do each becomes increasingly important. 

Communication still starts with understanding 

Communication is another capability that may become more important as AI makes producing communication easier. 

AI can draft an email, prepare a presentation, rewrite a policy or help a manager structure a difficult message. This can save time and help people communicate more clearly. But writing words and communicating well are not the same thing. 

Good communication begins before anything is written. It requires understanding what needs to be said, why it matters, who needs to hear it and how the message may be received. 

A manager can ask AI to draft a message about a change in working arrangements. The wording may be clear, professional and technically accurate. That does not mean the manager should simply send it. The manager still needs to understand what the change means for the team. They need to consider whether email is even the right way to communicate it. They need to think about what questions people are likely to have, what concerns may sit behind those questions and whether there is history or context that changes the way the message needs to be handled. 

Technology can help produce the message, but it cannot replace the human understanding of the relationships and context involved. 

As we explored in our recent article, “Different Expectations. Same Workplace.“, people can experience the same workplace differently. 

That matters in the context of AI because technology is becoming increasingly capable of recognizing patterns, personalizing information and responding to individual preferences. 

But personalization is not the same as understanding. 

AI may help us identify patterns in how people learn, communicate or engage at work. It can give leaders useful information and help them ask better questions. 

But information is only the starting point. Understanding what it means still requires context, conversation and judgment. 

And that is where the human capabilities we often call “soft skills” become even more important. 

While all of this information can be useful, personalization and understanding are not the same thing. Knowing more about a person does not automatically mean knowing what is needed in every situation. Someone who values autonomy may still need guidance. An employee who wants career progression may still need more time to develop. Preferences can provide valuable insight, but they do not replace leadership judgment. 

A person who prefers very little feedback may still need a direct conversation when performance concerns arise. Someone who values independence may need more support when expectations become unclear. An employee who is eager for a more senior role may still need an honest discussion about readiness, capability, and organizational needs. 

This is where leadership becomes more than simply responding to what people say they want. It requires leaders to listen, understand the broader context, and make thoughtful decisions about what is appropriate. Sometimes that means adapting. Sometimes it means explaining why something cannot be adapted. And sometimes it means making a decision that a person may not agree with while still helping them understand the reasons behind it. 

AI may help us know more about people. It cannot replace the responsibility to understand them. 

Empathy, adaptability and the human side of management

Empathy is often placed in the soft-skills category as well, and it is sometimes misunderstood as being nice, agreeable or avoiding difficult decisions. 

In a workplace context, empathy is something much more practical. It is the ability to understand how another person may experience a situation, even when the decision itself does not change. 

A leader may have to make a decision that is right for the organization but difficult for an employee. Understanding how that decision may affect the person does not mean reversing it. It means recognizing the impact and communicating with that reality in mind. 

This becomes particularly important as organizations use AI to support decisions about performance, development, workforce planning, and employee experience. 

A system may identify a pattern. A recommendation may appear logical. The available data may point in a particular direction. 

But decisions about people rarely exist only in the data. 

There may be context that was never captured in the system. There may be history that affects the situation. There may be practical consequences that are not visible in a dashboard. 

Empathy helps leaders see more of that picture. 

That is not separate from good decision-making. It is part of good decision-making. 

Adaptability is another capability likely to receive more attention as AI continues to change tasks, workflows, and expectations. People will need to learn new ways of working, some processes will evolve quickly, and certain roles may change significantly while others shift more gradually. 

Being adaptable will matter. But adaptability should not be confused with accepting every change simply because something new is possible. Real adaptability is not about chasing every new tool or trend. It is about being willing to learn, adjust, and let go of old ways of working when there is a good reason to do so, while also being able to assess whether a new approach is genuinely better. 

Sometimes that means moving quickly and embracing change. At other times, it means recognizing that an existing process is working well and does not need to be replaced simply because a new technology has become available. 

That distinction matters. Organizations do not need people who adopt technology uncritically. They need people who can evaluate it thoughtfully, understand where it adds value, and make sound decisions about when to use it and when not to. 

Once again, the differentiator is judgment. 

AI may remove some of the administrative work that currently takes up a manager’s time. It can summarize feedback, organize information, prepare first drafts, help structure meetings, and identify patterns that managers may want to explore further. All of this has the potential to make management more efficient. 

But greater efficiency does not necessarily make leadership easier. If AI prepares the information, managers still need to understand what it means. If AI drafts the communication, they still need to decide what should actually be said. And if AI identifies a potential performance issue, someone still has to determine whether it needs to be addressed and how that conversation should happen. 

As a result, the administrative side of management may become easier while the human side becomes more visible. Managers may spend less time gathering information and more time making sense of it. They may have fewer opportunities to avoid difficult conversations because they are busy with administrative tasks. Instead, they may be expected to create clarity, help people navigate uncertainty, and make decisions where there is no obvious answer. 

The value of a manager may also become less tied to simply knowing more than the people around them. Employees increasingly have access to powerful tools of their own. They can research information, compare options, generate ideas, and solve problems without always turning to a manager for answers. 

That does not make leadership less important. If anything, it changes what makes leadership valuable. A strong manager is not simply the person with the most information. They are the person who can help others make sense of that information, connect it to what matters, and turn it into meaningful action. 

Knowing how to use AI is not the same as knowing how to work well

Organizations are understandably investing in AI capability. Employees are learning how to use new tools, write better prompts, automate repetitive work, protect information, and understand the risks associated with AI. That training is important. 

The risk, however, is defining AI readiness too narrowly. Someone can become highly skilled at using AI without becoming equally skilled at evaluating what it produces. A manager can become more efficient without becoming more effective. A team can increase its output without improving the quality of its decisions. And an organization can introduce sophisticated technology while leaving the same underlying people and leadership challenges untouched. 

Unclear expectations do not become clear simply because AI is available. A manager who avoids difficult conversations does not become a stronger leader because they can generate a better draft. A weak process does not become a good process because it has been automated. In some cases, technology simply allows us to do an ineffective thing faster. 

That is why AI adoption is not only a technology challenge. It is also an organizational development challenge. The quality of leadership, culture, and expectations continues to matter because technology operates within that environment, not separately from it.  

Perhaps these skills were never soft 

The language organizations use influences what they pay attention to. When communication, empathy, critical thinking, adaptability, and judgment are described as soft skills, it becomes easy to treat them as secondary to technical expertise. 

Technical skills often feel easier to define and measure. Functional expertise is visible. Systems knowledge can be tested. Human capabilities are often harder to assess, but that does not make them less important. 

A technically strong manager who cannot provide clear feedback may still struggle to build a strong team. A highly capable specialist who cannot explain their thinking may struggle to influence decisions. And a leader with access to excellent data can still make poor decisions if their judgment is weak. 

Teams can also have all the right tools and still move confidently in the wrong direction if people do not feel comfortable challenging assumptions. These are not side issues. They are directly connected to performance. 

Perhaps AI gives organizations an opportunity to reconsider the language altogether. Rather than treating these capabilities as the softer side of work, we may need to recognize them for what they are: core human capabilities. They help us make sense of information, understand context, navigate relationships, make decisions, and take responsibility for what happens next. 

The real question is what we want people to become better at. AI will continue to change the skills people need. Organizations will need people who understand the technology, know how to use it responsibly, and can recognize where it creates real value. 

But technical AI capability is only part of the picture. The more technology can generate, the more important it becomes to know what is worth using. The more easily communication can be produced, the more important it becomes to understand what actually needs to be communicated. And the more systems can personalize an experience, the more important it becomes to understand the person behind the preference. 

The more AI can recommend, the more important it becomes for someone to exercise judgment. And the more work changes, the more important it becomes for leaders to help people make sense of that change. 

Organizations are still made up of people with different experiences, ambitions, expectations, and ways of seeing the world. Those people still need to work together toward shared goals. AI can expand what they are capable of doing, but it cannot decide what matters most, how people should work together, or what kind of organization they want to create. 

Those remain human questions. 

Perhaps that is why this is a good moment to rethink what we mean by soft skills. Maybe they were never soft at all. Maybe they are the capabilities that make everything else work.