Category: Uncategorized

Virtually every organization now has access to AI. Licenses have been purchased, tools rolled out, pilots launched. And yet, the results fall short of the promise: a handful of enthusiasts use the tools daily, while the rest rarely open them. This piece describes why AI adoption is fundamentally not a technology question but one of behavioral change, what skills employees need for it, and what organizations can learn from approaches that actually work in practice.

The tools are there, but the organization is searching

The figures on the investment side are unambiguous. The report Superagency in the Workplace by McKinsey (2025) shows that 92 percent of companies plan to invest more in AI over the coming years, while only 1 percent of leaders call their own organization mature in its usage. Even more striking is where the bottleneck lies according to that same study: not with the employees, who are further along than their executives think, but with leadership. Executives estimate that a small group of employees uses AI intensively, whereas the actual proportion is more than three times higher.

What I see in practice aligns with this. Organizations are flooded with AI solutions, and only then do the real questions begin. Is this allowed, and on what legal basis? Where is our data going? Who owns the risks? And above all: how do we get people to actually change their way of working? The technology has long been here. Trust and behavior are not always there yet.

Perhaps you recognize this: the licenses are arranged, the kick-off has taken place, and three months later everyone is working just as they did before.

What is AI adoption?

AI adoption is the process in which employees and teams genuinely and permanently integrate AI into their daily way of working. It is different from AI implementation: implementation is about access to tools; adoption is about different work behavior. An organization in which everyone has a license but no one adapts their working method has implemented, not adopted.

That distinction explains why so many AI initiatives stall after the pilot. Implementation is a project with an end date. Adoption is behavioral change, and behavior does not change through a rollout, but through repetition: trying something new, experiencing that it works, doing it again, until it becomes habit. Those who treat adoption as an IT project measure progress in rolled-out tools. Those who treat it as behavioral change measure progress in what people do differently.

New work demands new skills, and they aren’t technical

The skills that this adoption requires are only limitedly technical. Formulating a prompt can be learned by almost anyone in an afternoon. The Future of Jobs Report by the World Economic Forum (2025) shows which skills truly matter alongside technological literacy: critical and analytical thinking, curiosity and lifelong learning, resilience and flexibility. Translated into daily work: evaluating output instead of blindly accepting it, checking sources, recognizing bias, knowing when AI is precisely *not* the answer, and continuing to engage one’s own professional judgment. Those are soft skills, not button knowledge.

Since February 2025, this is no longer just an optional ambition either. Article 4 of the EU AI Act obliges organizations deploying AI to ensure a sufficient level of AI literacy among their employees, appropriate to their role and the context in which the systems are used. The question is therefore no longer *whether* employees should be developed in this area, but *how*: as a one-off e-learning course that fades after a week, or as practiced behavior that lasts.

What works in practice: five ingredients

A recognizable pattern emerges from the organizations where adoption does take off. Five ingredients consistently return.

First: governance in order, prior to the experiment. Clarity regarding privacy, data flows, and frameworks is not a brake but a prerequisite; it creates the trust to be allowed to experiment. Second: AI champions on the work floor, people who experiment themselves and bring colleagues along, up close and credibly. Third: a small learning rhythm instead of a big bang, for example learning one application every week and using it immediately. Fourth: AI in the real work context, not in a separate course environment; people adopt what makes their own work easier today. And fifth: room for job crafting. From the research by Wrzesniewski and Dutton (Academy of Management Review, 2001), we know that employees who actively redesign their own work experience greater ownership and meaning. Allowing people to reshape their work with AI, instead of imposing a new way of working, makes the change their own.

The common thread through all five: they are behavioral interventions. None of these ingredients are about the technology itself.

Practice as the engine of AI adoption

If AI adoption is behavioral change, the law of all behavioral change applies: knowing is not doing. A presentation on responsible AI usage changes work behavior just as little as a poster about feedback changes a culture. What changes behavior is practicing in one’s own work context, with immediate feedback, repeated over time.

That practice has two sides here. The first is the new work behavior itself: critically questioning AI output, testing advice from an AI system before building on it, explaining to a client or colleague how a decision was made. The second side is often forgotten: the conversations that support the change. The team leader who faces resistance from an employee who feels their profession is threatened. The AI champion who brings a skeptical colleague along without trying to persuade by pushing. The professional who explains to a concerned client what happens to their data. It is precisely those conversations that determine whether adoption shifts or stalls, and it is precisely those conversations that almost no one practices.

With AI Avatar training, that becomes possible: every conversation, scenario, and learning objective can be practiced in the employee’s own work context—from the resistance conversation to the critical assessment of AI advice—with immediate feedback on concrete behavior. This makes it visible, per competency and over time, whether the organization is truly moving forward; that is the difference between training and long-term retention. PractAIce is built on this exact idea, and is itself an example of AI that accelerates behavioral change rather than just taking over tasks.

What this means for HR and L&D

• Treat AI adoption as a behavior program, not as a rollout. Define what concrete work behavior you want to see six months from now, and design interventions around that: a small learning rhythm, practice in the work context, champions close at hand.
• Develop AI literacy as a skill, not as a certificate. Article 4 of the EU AI Act calls for a sufficient level appropriate to the role. A ticked-off e-learning course demonstrates participation; repeated and evaluated behavior demonstrates literacy.
• Train the conversations surrounding the change as well. Managers and AI champions carry the adoption. Let them practice dealing with resistance, concerns, and bringing colleagues along before those conversations really matter.

Frequently Asked Questions

What is AI adoption?

AI adoption is the process in which employees and teams genuinely and permanently integrate AI into their daily way of working. It differs from implementation: implementation is about access to tools, whereas adoption is about lasting changes in work behavior.

What is AI literacy and is it mandatory?

AI literacy is the ability to use AI systems in an informed manner: evaluating output, recognizing risks and bias, and knowing where the boundaries lie. Since February 2025, Article 4 of the EU AI Act obliges organizations that use AI to ensure a sufficient level of AI literacy among their employees, appropriate to their role.

How do you get employees onboard with AI?

Not with a big launch, but with behavioral interventions: clear frameworks that instill trust, champions on the work floor, a small weekly learning rhythm, application in their own work, and room to reshape the work itself. Practicing with immediate feedback, including on the conversations surrounding the change, makes the difference between a pilot and lasting behavioral change.

About the author: Sven is the founder of PractAIce and a behavioral expert. He has worked for years on driving behavioral change, embedding new habits into daily practice, and making soft skills concrete and measurable.

Wondering if AI adoption in your organization is moving beyond the pilot stage? With PractAIce, employees and managers practice the behaviors and conversations that drive change, in their own work context, with immediate feedback. Discover AI Avatar training or request a demo today.

Few professional groups feel the rise of AI as directly as trainers and coaches. E-learning platforms promise the same results for a fraction of the price, clients demand demonstrable impact, and the traditional one-day training course is losing ground. The natural knee-jerk reaction is understandable: viewing AI as a competitor. This article outlines why that is the wrong interpretation, and how AI role-plays can instead become the missing piece of your value proposition: the practice arena between sessions, delivering proven behavioral change as the end result.

The environment for trainers is changing faster than their offerings

The demand for personal and professional development has never been higher. The Future of Jobs Report by the World Economic Forum (2025) shows that employers expect 39 percent of workers’ core skills to change by 2030. According to research by Gartner, 85 percent of business leaders expect the need for skill development to surge in the coming years due to AI and digitalization, with many feeling that traditional learning reacts too slowly to the pace of this shift.

At the same time, what organizations are asking for is changing. The Workplace Learning Report by LinkedIn (2025) reveals that nearly half of L&D professionals hear executives question whether employees possess the skills needed to execute company strategy. Organizations must become more agile, shifting the demand from generic programs toward tailored training: focused on the specific conversations, clients, and tensions of *this* team, with visible impact on behavior. Clients are asking very different questions than they did five years ago. No longer “what does the training cover?”, but “what actually changes on the work floor afterward, and how do we prove it?”

You might recognize this shift in your own proposals. The content still convinces, but the demand for retention and evidence grows stronger with every pitch.

The gap in the traditional proposition

This is precisely where the classic training model falls short. From the research on deliberate practice by K. Anders Ericsson (Psychological Review, 1993), we know that true skill mastery comes from targeted, repeated practice paired with immediate feedback. A single training day can deliver insight and initial practice, but it cannot offer repetition: a trainer cannot stand beside every participant for weeks as they apply what they learned. This creates the well-known transfer problem: what was learned in the classroom fades in the weeks that follow, and no one can prove what actually stuck.

The gap in many training companies’ value proposition isn’t the quality of the training itself, but what happens *after* the session. Whoever bridges that gap fundamentally strengthens their position.

AI role-plays: an extension of the trainer, not a replacement

The productive view of AI role-playing isn’t replacement—it’s extension. The trainer remains the expert who reads the group, sets the methodology, and facilitates the real human conversation. What AI adds is scale—something one human simply cannot provide: every participant can keep practicing between sessions, at their own convenience, in scenarios tailored to their daily work.

In an AI role-play, via AI Avatar training, a participant re-enacts the scenario from the workshop with an AI counterpart that responds realistically: offering pushback, expressing emotion, and adapting where a static script would fail. Every session provides immediate feedback on concrete behavior. The trainer then sees not only *that* participants practiced, but *how* the group is progressing, competency by competency over time.

PractAIce as a strategic tool for training firms

PractAIce was built to fulfill this role, keeping the trainer as the true owner of the content. Trainers and training companies build their own AI role-plays and embed their own methodology and frameworks into the feedback system, ensuring the AI evaluates along the exact principles the trainer teaches. Every conversation, scenario, learning objective, and skill can be trained within the participant’s actual work context—whether it’s the sales call for a specific sales team or the difficult news conversation for a healthcare provider.

In this way, AI ceases to be an expense or a threat; it becomes part of your business model. Strategically, this transforms the training firm’s position. The proposition shifts from selling workshop days to delivering end-to-end programs with measurable results: workshop, practice phase, and progress measurement forming a unified whole. This sets you apart in tenders where long-term retention and measurability are increasingly required, creates a recurring revenue stream alongside training sessions, and answers the demand of organizations looking for agile, customized solutions. The trainer is no longer a reseller of technology, but uses the platform as a direct extension of their own craft.

How a training agency puts this into practice

A concrete example makes this clear. A training agency delivers a leadership program on constructive feedback and tough conversations to a mid-sized healthcare provider. Previously, this consisted of two training days and a follow-up half-day session. With PractAIce as a tool, the program looks fundamentally different.

The agency translates the client intake into role-plays written in the client’s language: a conversation with an overworked nurse, or addressing a colleague regarding protocol compliance. Following the first training day, participants spend six weeks doing short, twice-weekly AI role-plays with an AI Avatar, scored against the agency’s custom feedback model. The trainer tracks group-level progress and spots where people get stuck—for instance, digging deeper after encountering resistance. As a result, the follow-up session isn’t a recap of theory from six weeks ago, but a targeted deep-dive into patterns revealed by the practice data. The client receives a clear report showing competency development over time.

Three things change for the agency simultaneously:
1. The trainer shifts from a lecturer to a data-informed coach.
2. The client sees black-and-white proof of the program’s return on investment, making contract renewals significantly easier.
3. The agency gains a recurring platform license component in its offerings without ever having to build software themselves.

What this means for trainers and coaches

• Reposition your offer around behavior and evidence. The market is shifting from knowledge transfer to demonstrable behavioral change. Structuring your programs around practice journeys backed by analytics aligns directly with client demands.
• Use AI as an extension of your craft, not as a gimmick. The true value lies not in the technology itself, but in the repeated practice and behavioral insights it places directly into the trainer’s hands.
• Choose a platform where your own content takes center stage. A trainer who can build their own methodology, scenarios, and feedback frameworks reinforces their unique value proposition. A rigid, closed library makes you interchangeable.

Frequently Asked Questions

What is an AI role-play?
An AI role-play is a conversational practice exercise where a participant engages in a realistic dialogue with an AI Avatar that responds like a real human: complete with resistance, emotion, and distinct objectives. Afterward, the participant receives feedback on specific behaviors, enabling repeated practice to translate into visible development.

Does AI replace the trainer?
No. The trainer sets the methodology, reads the room, and leads the high-value conversations that only a human can facilitate. AI takes over the component that is practically impossible for a human trainer: providing weeks of continuous practice with instant, individual feedback. In return, the trainer receives behavioral data to sharpen their own coaching.

How do you integrate AI into an existing training program?
By structuring the practice period between touchpoints: translate the program’s learning goals into role-play scenarios set in the participants’ real work context, have them engage in short, frequent practice sessions between workshops, and use the progress data to guide the follow-up session. This ensures that training and practice reinforce each other rather than existing in silos.

---

About the author: Sven is the founder of PractAIce and a behavioral expert. He has worked for years on driving behavioral change, embedding new habits into daily practice, and making soft skills concrete and measurable.

Wondering if this fits your training practice? PractAIce is built by and for trainers and coaches: you build your own role-plays, maintain your own methodology, and gain the behavioral data that elevates your proposition. Discover the possibilities or request a demo today.

Almost every organization wants a feedback-and-accountability culture (*aanspreekcultuur*). The core values have been defined, the feedback training has been completed, and the model is figuratively hanging on the wall. And yet, the conversation still doesn’t happen: the colleague who consistently delivers late is not called out, concerns about workload are swallowed, and a mistake is only discussed when it can no longer be hidden. This article describes what a feedback culture truly is, why giving feedback goes wrong so often, and why practiced behavior—rather than knowledge of models—determines whether people actually speak up to one another.

The values are on paper, but the conversation remains unheld

Most organizations have carefully articulated their values: ownership, respect, collaboration, openness. Often, the specific behavior associated with them has even been made concrete. The challenge rarely lies in defining those values, but in consistently applying them in daily work. Because that is precisely where differences arise. To one person, ownership means taking initiative; to another, it means honoring agreements. To one person, respect means giving space; to another, it means having the honest conversation.

Culture does not change because values are communicated. Culture changes when people recognize, practice, discuss, and repeat the behavior behind those values. However, the hustle of the day quickly shifts focus back to deadlines and operational priorities. If speaking up is not given a fixed place in daily work, that culture remains something people talk about, rather than something people do.

Perhaps you recognize this: the training has been completed, everyone knows the feedback rules, and in the very next team meeting, nobody says a word again.

 

What is a feedback culture?

A feedback culture is a work environment where colleagues address each other on behavior, agreements, and results in a timely, direct, and respectful manner, and where giving and receiving feedback is a normal part of daily work rather than an exception. It is not about harshness, but about clarity: issues are discussed while they are still small, directly with the person involved, and with the intention of improving together.

That final point distinguishes speaking up from blaming or taking someone to task. Addressing someone is focused on behavior and the future; calling someone out or punishing them is focused on the person and the past. The difference between the two lies not in the intention on paper, but in how the conversation is actually conducted. And that is precisely where things often go wrong.

Why giving feedback goes wrong so often

Feedback is often treated as if more is always better. Research shows a more uncomfortable picture. In the classic meta-analysis by Kluger and DeNisi (Psychological Bulletin, 1996), based on hundreds of studied feedback interventions,

more than a third of the interventions actually reduced performance. Feedback that shifts attention toward the person rather than the task triggers defensiveness instead of learning. Hattie and Timperley (Review of Educational Research, 2007) reach a similar conclusion: feedback is among the most powerful learning interventions we know, but only when it is aimed at the task, the process, and the recipient’s self-regulation, and not at the person.

Research by Gallup confirms that the problem lies not in the need for feedback, but in its execution: employees who receive valuable feedback are far more engaged and significantly less burned out, but a large portion of employees are not looking for more frequent feedback because the feedback they do receive is not valuable.

The conclusion can be sharply defined. Giving feedback is not a knowledge problem. Almost anyone can list the rules: describe the behavior, state the effect, do it in a timely manner. The problem lies in execution under pressure: the tone in the very first sentence, the silence after the message, following up with questions when the other person becomes defensive. That is behavior, and behavior is not learned from a model.

Psychological safety as a prerequisite

There is a second layer as well. Speaking up requires that people feel safe enough to take the risk of having the conversation. Research by Amy Edmondson (Administrative Science Quarterly, 1999) showed that teams with high psychological safety report mistakes sooner and learn faster; not because more things go wrong, but because what goes wrong is actually discussed.

That safety is not a policy document either, but the sum of behaviors. It arises in a manager’s reaction to the first critical remark, and in the way a team deals with a colleague who shares a mistake. A single punitive reaction does more damage to that culture than ten posters can build up. Anyone who wants to strengthen the culture must therefore start with this behavior, and especially with the behavior of leaders.

You practice a feedback culture, conversation by conversation

Knowing is not doing. Between understanding a feedback model and conducting a good conversation lies practice: trying, reviewing, trying again. The problem has always been that this space to practice was missing. Practicing on a actual colleague is risky, and hiring an actor for every employee is not scalable.

This is where AI substantially changes the possibilities. With AI avatars that respond like real people, every employee and manager can practice the exact conversations they need to hold in their own daily work: addressing a colleague about an unfulfilled agreement, making workload open for discussion, sharing a concern about quality or safety, or receiving feedback without becoming defensive. Every conversation, scenario, and learning objective can be practiced in one’s own work context. After practicing, targeted feedback and feedforward follow: what went well, what could be stronger, and what alternative behavior would help in the next situation. The AI coach helps identify patterns and make targeted adjustments to personal development.

Additionally, teams can create, share, and discuss roleplay scenarios themselves. This generates not only individual growth, but a shared language for addressing one another: what does it sound like here when someone brings up an agreement? That is the difference between a single training and true embedding: not practiced once, but repeated, recognizable, and scalable. In this way, a feedback culture becomes something that people build together, conversation by conversation. PractAIce was built for precisely that movement.

What this means for HR and leaders

• Start with leaders. Their response to criticism and mistakes determines the safety of the entire team. Have them be the first to practice receiving feedback, not just giving it.
• Make practicing a permanent part of work. A single feedback training will not change a culture. Short, repeated practice moments built around real situations from daily work will.
• Measure behavior, not satisfaction. A high satisfaction rating for a training says nothing about whether the conversation actually takes place afterward. Look at observable behavior over time: are people speaking up sooner, more often, and better?

Frequently Asked Questions

What is a feedback culture?

A feedback culture is a work environment where colleagues address each other on behavior, agreements, and results in a timely, direct, and respectful manner. Giving and receiving feedback is a normal part of daily work, aimed at improving together rather than placing blame.

How do you create a feedback culture in your team?

By practicing behavior instead of merely communicating values. Start with the manager, make desired feedback behaviors concrete, have people repeatedly practice them in recognizable situations from their own work, and discuss as a team what works. Psychological safety grows with every positive experience.

What is the difference between feedback and feedforward?

Feedback looks back: what did you do, and what effect did that have? Feedforward looks forward: what concrete alternative behavior will help in the next, similar situation? The combination works best, because the recipient not only hears what happened, but also knows what they can do differently tomorrow.

About the author: Sven is the founder of PractAIce and a behavioral expert. He has worked for years on behavioral change, embedding new habits into daily practice, and making soft skills concrete and measurable.

Are you unsure whether speaking up in your organization is moving from paper to practice? With PractAIce, employees and leaders practice giving and receiving feedback with an AI Avatar in their own work context, supported by immediate feedback and feedforward. Discover AI Avatar training or feel free to request a demo.

Ask an executive about the biggest risks, and the answer will focus on the market, technology, or finance—rarely on skills. Yet, a risk is quietly mounting in precisely that area: the skills gap—the disparity between people’s current capabilities and the demands of the work. This article explains why that gap is widening faster than ever, how ignored gaps accumulate into “skills debt,” and why practice, feedback, and repetition offer the only reliable path to bridging the divide.

The work changes, the job description does not

The figures on the labor market have been pointing in the same direction for years. According to research by Gartner, 58 percent of the workforce needs new skills to continue performing their jobs successfully. The number of skills required for a single role has been increasing by around ten percent per year since 2017, and one in three skills listed in a 2017 job posting has now become outdated. The Future of Jobs Report by the World Economic Forum (2025) adds a forward-looking perspective: employers expect that 39 percent of employees’ core skills will change by 2030.

Compare those figures with the average organization, and the tension becomes clear. Job architectures are revised only every few years. Learning and development plans are organized by calendar year. Performance review cycles look backward. The work itself pays no attention to any of that: tasks shift from month to month, and AI changes work processes while they are already in motion. On paper, every position is filled. In practice, the work beneath those positions has already changed. Roles are static, skills are dynamic, and it is precisely in that gap that the skills gap emerges.

You may recognize the situation. The vacancy has been filled, the team is complete, and yet the question remains: are the people in those roles truly equipped for the work that awaits them next year?

What exactly is a skills gap?

A skills gap is the difference between the skills employees currently possess and the skills required for the work—both today and in the near future. While this may sound abstract, it can be made concrete by asking three questions: What skills does the work require, both now and two years from now? What skills are already present in-house—often exceeding what the formal job structure indicates? And which skills are missing? That third question is where targeted development begins.

Anyone who seriously addresses these questions usually discovers two things. First, the organization possesses more capabilities than job titles suggest: people perform work that is not documented anywhere. Second, skill shortages rarely align with where training budgets are currently allocated. This lies at the heart of a skills-based approach: focusing on skills—rather than job roles—as the fundamental unit for assessment, development, and deployment.

From Skill Gap to Skill Debt

Software developers are familiar with the concept of technical debt: every quick fix you choose today becomes maintenance you will eventually have to perform later—with interest. The same mechanism applies to skills. A skill gap that is visible today can be measured and addressed. A gap that is left unattended becomes skill debt: deferred maintenance on an organization’s ability to do its work.

This debt does not appear in a single report but creeps in gradually. In customer conversations that become more difficult. In projects that slow down because no one has truly mastered the new way of working. In employees leaving because they are no longer developing. Research by Deloitte shows the other side of the same coin: organizations that focus on skills deploy talent far more effectively and anticipate change better than organizations that continue to rely on traditional job-based thinking. Organizations that manage by skills gain an advantage—but only if they are able to develop those skills.

Why Soft Skills Represent the Biggest Gap

Not every gap is equally visible. A shortage of technical skills is easy to spot: a missing certification or a system that no one knows how to operate. The gap in soft skills is of a different nature. As AI takes over more task-based work, the value of human work shifts toward what cannot be automated: collaboration, critical thinking, leadership, and the ability to adapt while the work itself continues to change. The World Economic Forum consistently ranks these skills among the most important core skills for 2030.

At the same time, these are the skills that are the hardest to measure. “Handles a difficult conversation calmly and clearly” does not appear on any diploma. It is behavior, and behavior only reveals itself in real situations. This creates an uncomfortable reality: the skills that will become most critical to organizations are also the skills that no one knows with certainty who truly possesses. As a result, the largest gap is also the least visible.

Closing the Gap: Practice, Feedback, Repeat

The standard response to identifying a gap is predictable: purchase a training course. But transferring knowledge does not close a behavioral gap. Ever since Hermann Ebbinghaus described the forgetting curve—replicated in 2015 by Murre and Dros—we have known that most newly acquired knowledge fades within days when it is not reinforced through repetition. A skill is not developed by hearing about it, but by doing it: practicing, receiving immediate feedback, trying again, and applying it in real work.

For a long time, this has been the practical challenge: practice with feedback is labor-intensive and difficult to scale. AI role-playing changes that equation. With an AI Avatar training, employees practice the conversation they will actually have tomorrow, within their own work context, with an AI counterpart that responds realistically. Every scenario, every learning objective, and every skill can be practiced, and every session provides feedback on concrete behavior. This not only closes the gap but also makes it visible: by competency, by team, and over time. Learning shifts from an annual HR activity to continuous infrastructure for organizational agility. PractAIce is built on that principle.

What This Means for HR and L&D

Three key implications stand out:

• Map the gap by skill, not by job title. Do not ask which positions are vacant, but which behaviors the work of the next two years will require and who can demonstrably perform them today. The answer is often very different from what the organizational chart suggests.
• Treat the gap as an ongoing challenge. An annual learning plan fits a labor market that changes once a year, not work that shifts every month. Short, frequent practice sessions with feedback keep skills up to date in ways that occasional training never can.
• Measure progress by behavior, not participation. Completing a course only shows that someone attended. Only repeated, observable, and assessed behavior demonstrates that the gap is actually becoming smaller.

One important nuance should be added. Behavioral data are development data, not a performance evaluation tool. Their value lies in making growth visible. Organizations that safeguard this distinction build the trust people need to practice and improve their behavior.

Frequently Asked Questions

What is a skill gap?

A skill gap is the difference between the skills employees currently possess and the skills that the work requires, both now and in the near future. The gap can be identified for each team and each skill and serves as the starting point for targeted development.

What is skill debt?

Skill debt is the accumulated impact of skill gaps that remain unaddressed: deferred maintenance on an organization’s capabilities. Like technical debt, skill debt grows quietly over time until it becomes visible in quality, agility, and employee turnover.

How do you close a skill gap in your organization?

Start by measuring: what behaviors does the work require, and who can demonstrably perform them? Then develop those behaviors through repeated practice with immediate feedback, preferably within the employee’s own work context. AI role-playing makes this type of practice scalable and enables measurable progress for every competency.

About the author: Sven is the founder of PractAIce and a behavioral change expert. For many years, he has focused on helping organizations achieve lasting behavioral change, embedding new behaviors into daily practice, and making soft skills concrete and measurable.

Not sure where the biggest gap in your organization lies? PractAIce makes skills visible and trainable. Employees practice every conversation, scenario, and learning objective with an

AI Avatar in their own work context, while their development becomes measurable for every competency. Discover the AI Avatar training or request a demo.

The average training course exists, but the average employee does not. Yet everyone receives the same schedule, the same pace, and the same examples—and we simply hope it fits. For the first time, AI makes something different possible: learning that adapts to the individual. This article explores hyper-personalized learning, why a one-size-fits-all approach falls short—especially regarding soft skills—and how an AI avatar facilitates tailored practice for everyone’s personal development.

The Tyranny of the Average

A traditional soft skills training course is, by necessity, designed for the middle ground. The experienced salesperson and the beginner sit in the same room, receive the same instruction, and practice the same case study. For one, the pace is too fast; for the other, too slow; and for most, the scenario is close—but not quite relevant to their own situation. This is not a criticism of the trainer; it is a limitation of the format. One person simply cannot effectively support twenty people at their individual skill levels simultaneously.

The consequences are well known. People disengage because the training does not feel relevant, or they leave with lessons that are general enough to be useful to no one in particular. Soft skills suffer the most from this because they are highly dependent on context. Communicating effectively with an angry customer requires a different approach than addressing a colleague about their behavior, which in turn is different from having a difficult conversation with an employee. A single generic exercise cannot cover all of those differences.

Why a tailored approach matters—especially for soft skills

There is a long-standing finding in educational science that highlights this issue. In 1984, educational psychologist Benjamin Bloom described what he called the “2 Sigma Problem” in the journal *Educational Researcher*: students receiving one-on-one tutoring performed dramatically better than those in a standard classroom. Personalized guidance is thus demonstrably effective. However, it was never scalable; you cannot assign every employee a dedicated coach who is available at all times. Consequently, personalized learning has historically been a privilege rather than the norm.

This is particularly significant when it comes to soft skills, as development in this area relies on practicing at precisely the right level. If the material is too easy, no progress is made; if it is too difficult, people disengage. Research on “deliberate practice” by K. Anders Ericsson (*Psychological Review*, 1993) demonstrates that growth stems from practicing just outside one’s comfort zone, accompanied by immediate feedback. That “just outside” zone varies from person to person—which is precisely what standardized training cannot accommodate.

This explains why so many training programs never rise above the level of being merely “reasonably useful.” Those who are already proficient become bored, while beginners feel overwhelmed. Genuine progress requires the level of difficulty to adapt to the individual: raising the bar when a task is mastered and lowering it slightly when someone gets stuck. It is this very adaptability—impossible to achieve with fixed case studies and a set pace—that makes personalized learning so powerful.

What hyper-personalized learning is

Hyper-personalized learning is a form of development that adapts to the individual employee—fitting their proficiency level, role, goals, and pace. Instead of putting everyone through the same material, what a person already knows and what they need determines what they practice. It is adaptive, tailored, and focused on situations relevant to that specific individual. It represents the difference between training you simply undergo and development built around you.

The AI Avatar as a Personal Sparring Partner

What remained out of reach for Bloom—personalized guidance at scale—is now becoming achievable with AI. An AI role-play effectively gives every employee their own personal sparring partner, available whenever they need it. The AI Avatar adapts to the individual: it conducts conversations tailored to each person’s role and skill level, increases the level of difficulty as they progress, and selects scenarios that are relevant to that specific person, such as dealing with a difficult customer, handling a resistant colleague, or delivering bad news.

The feedback is equally personalized. It is not a generic list of points for improvement, but specific feedback on what the individual just did: where they moved too fast, where they failed to ask probing questions, or where they lost the other person’s attention. As a result, no one is merely practicing ‘communication in general’; instead, they are honing the exact skill they wish to improve. In this way, AI training and AI avatar training transform personal development into something unique—a process that takes a different form for every employee.

Two people, the same platform, yet vastly different learning paths. An experienced account manager might practice closing a difficult deal with a tough, inquisitive counterpart. A novice, meanwhile, might first practice calmly opening that same conversation without freezing up. They use the same AI avatar but receive a scenario, difficulty level, and feedback tailored to their specific development needs. In a traditional training session, they would share the same case study; here, no two people share the same one.

It also scales in ways a human coach cannot. Twenty people can practice simultaneously—each focusing on their own skills at their own level—while their progress is tracked for each specific competency. This is where the approach aligns with competency-based training: rather than putting the entire group through the same course, each individual follows their own developmental trajectory. Consequently, role-playing exercises become more personalized than ever, rather than more generic.

Personalization without replacing the human element

A valid objection is that personal development is, after all, a fundamentally human endeavor. That is true, and hyper-personalized learning does not deny it—it simply distributes the work differently. AI takes over the aspects that do not scale well: endless repetition, tailored practice, and immediate feedback on behavior. This frees up time and attention for the part that is truly human.

A coach or manager no longer needs to oversee twenty identical practice sessions; instead, they can engage in the conversation that really matters—discussing patterns, meaning, and the personal step an individual needs to take. Technology does not render the human element obsolete; rather, it restores the most valuable role a person can play.

What this requires of organizations

Hyper-personalized learning primarily requires the courage to let go of the average. Three principles help:

• Start with the individual, not the course. Let people practice in situations and at levels relevant to *them*, rather than subjecting everyone to the same program. That is competency-based training in practice: focusing on skills rather than the curriculum.
• Provide opportunities to practice, not just to listen. Personal growth stems from hands-on experience at the right level—accompanied by feedback—rather than from a generic presentation.
• Use individual progress as your compass. Tailor development to what each person demonstrates and needs, and make this the foundation of the development conversation.

The associated concern is valid: doesn’t the use of so much personal data turn people into mere objects of measurement? The answer lies in the purpose. This data serves the employee’s development, not their surveillance. When implemented effectively, hyper-personalized learning actually gives people greater control—over what they practice, the pace, and the goals they work towards.

Ultimately, the promise is simple. It is not about funneling everyone through the same process and hoping for a fit, but about developing each individual in the way that works best for them. For a long time, this was prohibitively expensive. Now, for the first time, it is simply possible.

Frequently Asked Questions

What Is Hyper-Personalized Learning?

Hyper-personalized learning is an approach to development that adapts to each individual employee—their skill level, role, goals, and pace. Instead of offering a single program for everyone, it determines what and how someone practices based on what they already know and what they need to develop.

How Does AI Personalize Learning?

AI personalizes learning by adapting to the individual user. An AI role-play selects scenarios that match a person’s role, increases the level of difficulty as they improve, and provides feedback based on their own behavior rather than offering generic advice.

Does AI Replace the Trainer or Coach?

No. AI takes care of repetitive practice and provides immediate feedback at a scale that is impossible for a human to achieve. The trainer or coach retains the uniquely human role: providing direction, adding meaning, and facilitating the conversations that require deeper reflection.

About the Author — Sven is the founder of PractAIce and a behavioral expert. For many years, he has focused on helping people change their behavior, embedding new habits into daily practice, and making soft skills development concrete and measurable.

PractAIce makes personalized development at scale possible. Every employee practices with an AI Avatar that adapts to their skill level and goals. Would you like to see what this looks like for your team? Discover AI Avatar Training or schedule a demo.

A training day feels like learning but plays out like an event: a single moment, a burst of energy, and then silence. Yet skills aren’t built in a single day; they develop through repetition in the actual work. This piece explores this new approach to learning—learning in the flow of work, spaced out and practiced—and explains why AI role-playing finally makes this shift practical for soft skills.

The problem with the training day

The classic approach is well known: a day off the schedule, an external trainer, a conference room, and a satisfaction score at the end. The problem lies not in the day’s content, but in its format. Learning is treated as an event with a clear beginning and end, whereas behavior changes based on what happens *after* that day. As early as 1885, Hermann Ebbinghaus demonstrated via his “forgetting curve” just how quickly knowledge fades without repetition; many representations of his work show that a significant portion is lost within a week. A one-off training day, therefore, fights against the very nature of human memory.

Then there is the structure of learning itself. The well-known 70-20-10 model—popularized by Charles Jennings, among others—estimates that people acquire roughly ten percent of their knowledge through formal training, twenty percent from and with others, and seventy percent by doing the work itself. If this holds true, many organizations are investing the lion’s share of their training budgets into the smallest component of learning. A training day is not worthless, but it is overrated when it stands alone.

There is also a blind spot regarding how we measure training success. A satisfaction score at the end of the day reflects the experience, not behavior a month later. People leave feeling enthusiastic, yet the old patterns often return within weeks. This is not because the day was poor, but simply because a one-time spike does not create a habit. Anyone serious about learning measures not whether it was enjoyable, but whether it actually stuck.

What ‘new learning’ means

The new approach to learning shifts the focus from the event to the process. Instead of a single training day each year, it creates a continuous flow of short learning moments woven into the work itself—an approach referred to in the literature as learning in the flow of work. Learning is no longer an interruption to work, but an integral part of it: brief, relevant at the moment it matters, and repeated often enough to stick.

For knowledge acquisition, this has already become common practice. We look up an explanation when we need it, watch a short video, or read an article. But for soft skills—such as communication, giving feedback, leadership, and handling resistance—simply looking something up is not enough. You do not learn how to handle a difficult conversation by reading about it; you learn by having the conversation. Until recently, practicing these kinds of behaviors was difficult to integrate into the flow of work. That is exactly what is changing now.

What the new way of learning is not

“New learning” is often confused with digitization. Recording a classroom training session and offering it as a video module, or lining up a series of e-learning courses, feels modern but changes little at the core. It remains a matter of knowledge transfer—just on a screen instead of in a classroom. The employee watches, clicks through, and forgets just as quickly as they would in a live session.

The difference lies not in the medium, but in the learning method. New learning is about repeatedly practicing behavior and receiving feedback, not about consuming content. Watching a video is not practice; holding a conversation is. Confusing the two means digitizing the old problem rather than solving it.

Soft skills: practice, don’t just acquire knowledge

The science regarding skill acquisition is unequivocal on this point. Research into “deliberate practice” by K. Anders Ericsson (*Psychological Review*, 1993) shows that expertise stems not from experience per se, but from focused, repetitive practice involving immediate feedback and increasing levels of difficulty. A pianist improves not through a single masterclass, but through daily, focused practice. The same applies to conversational skills—an uncomfortable truth for one-day training courses: without repetition and feedback, little actually changes.

Traditional role-playing acknowledged this—practicing with an actor or colleague is more valuable than listening to theory. However, it does not scale well. Actors are expensive and in short supply; the exercise usually takes place only once; and many people find practicing in front of a group uncomfortable. Consequently, the experience remained an isolated highlight rather than becoming a habit. The repetition required for genuine learning was missing.

AI role-playing makes the new way of learning practical.

This is where technology bridges the gap. With an AI role-play, someone practices a conversation with an AI Avatar that responds realistically: becoming curt, pushing back, and adapting in ways that go far beyond a scripted interaction. Because these practice sessions can take place anytime, without anyone watching, and as often as needed, they fit naturally into the flow of work in a way that a one-day training course never could. Just ten minutes between two meetings is enough for a valuable practice session.

The impact is tangible. A sales team that replaces its annual training day with one short practice session each week does not practice once per quarter, but ten times—each conversation slightly different, with immediate feedback. It is not the intensity of a single day that drives improvement, but the cumulative effect of many small learning moments. This is exactly the shift predicted by both the forgetting curve and the principles of deliberate practice.

Practice does not stop with the conversation itself. After every session, participants receive feedback on specific behaviors—such as conversation structure and asking follow-up questions—linked directly to what they just did. This creates the combination that deliberate practice requires: repetition, progressively increasing challenge, and immediate feedback. AI training, AI Avatar training, and role-playing exercises are transformed from one-off workshops into an ongoing practice routine. This is also where reinforcement finally becomes a natural part of the process: embedded in a consistent learning rhythm.

In addition, every practice session is evaluated against individual competencies. This makes the new approach to learning a natural fit for competency-based training: the focus shifts from completing courses to developing skills that can be measured and tracked over time. Personal development becomes concrete and meaningful, rather than simply another box checked on a training attendance list.

What This Requires from Organizations

The new approach to learning is less about purchasing another training program and more about adopting a different habit: shifting from standalone soft skills training to an ongoing practice routine. It aligns naturally with competency-based training, where the focus is on developing skills rather than completing courses. A few key shifts make the difference:

• Shift from events to rhythm. Replace the ambition of one major training day with short, frequent practice sessions. It is not about learning less, but about distributing learning differently.
• Make practice the norm. People only practice new behaviors when it feels safe to make mistakes. An environment without an audience or judgment lowers that barrier and encourages experimentation.
• Focus on observable behavior. Measure development by what people demonstrate rather than by attendance, and use competency-based feedback as the foundation for coaching conversations.

This does not make the trainer obsolete. Instead, the trainer’s role shifts from delivering knowledge to designing and facilitating meaningful practice. They provide direction, add context, and lead the conversations that AI cannot. The repetitive practice is automated, while the human contribution becomes even more valuable.

Perhaps the greatest benefit is that this new approach solves an old problem. The gap between knowing and doing, between the training room and the workplace, was never caused by poor training. It was a matter of format. By embedding learning into everyday work and making behavioral practice a continuous part of the job, that gap finally begins to close.

Frequently Asked Questions

What is the new approach to learning?

The new approach to learning shifts the focus from one-off training days to continuous learning embedded in everyday work—short, relevant, and repeated over time. Instead of transferring knowledge in a single event, it emphasizes practicing skills consistently so they are retained and applied more effectively.

What is learning in the flow of work?

Learning in the flow of work means that learning becomes part of daily work rather than an interruption to it. People learn or practice when it is most relevant, through short sessions, without having to set aside an entire day for a training course.

How do you practice soft skills digitally?

Soft skills can be practiced digitally through AI role-playing. You engage in realistic conversations with an AI Avatar that responds like a real conversation partner, followed by feedback on your behavior. This allows you to practice communication, giving feedback, leadership, and other interpersonal skills as often as needed, over time.

About the author — Sven is the founder of PractAIce and a behavioral change expert. For many years, he has focused on helping organizations achieve lasting behavioral change, embedding new behaviors into everyday practice, and making soft skills development tangible and measurable.

PractAIce puts the new approach to learning into practice. Employees engage in realistic conversations with an AI Avatar whenever they need, receiving immediate feedback after every session. Would you like to see how this works in your organization? Discover AI Avatar Training or schedule a demo.

For decades, the diploma was the currency used to express talent. That currency is losing value. It is not the skill itself, but the proof of mastery, that is becoming the new benchmark. This article outlines the shift toward the skills-based organization, explains why soft skills are the hardest to capture in this context, and describes how a skills passport—populated with demonstrably practiced behavior—can bridge that gap.

The erosion of the diploma as a benchmark

A diploma certifies a moment in time. It shows that someone met certain requirements at a specific point in time—not that they still possess the skills their job requires today. For a long time, that distinction was largely academic. Not anymore. In the Future of Jobs Report by the World Economic Forum (2023), employers expected that 44 percent of the skills required for work would change within five years, and that six in ten employees would need reskilling before 2027. A widely cited estimate suggests that the lifespan of a skill has now shrunk to around five years—and even less for technical knowledge. What people once learned becomes outdated faster than ever before.

The result is a quiet revaluation of what truly matters. More and more employers are replacing formal degree requirements with demonstrable skills—skills-based hiring—and are organizing their businesses around the skills the work requires rather than around fixed job roles. Research firm Deloitte

describes this shift as the “skills-based organization”: a model in which the skill, rather than the job title, becomes the unit of work. Underlying research reveals just how far practice has already shifted—63 percent of work now falls outside formal job descriptions. What a person can actually do carries more weight than the specific box they fill on an organizational chart.

Major employers and government bodies—ranging from tech companies to public sector organizations—are removing degree requirements from job postings and selecting candidates based on their capabilities rather than where they studied. At the same time, this movement is still in its infancy: fewer than one in five organizations has truly made the transition to a skills-based approach. The underlying realization is simple: a degree is becoming an increasingly poor predictor of whether someone can handle today’s work. Consequently, experienced professionals lacking the “right” credentials are becoming visible again, while impressive CVs that lack up-to-date skills are losing their value.

For learning and development, this represents a fundamental shift. A training budget evaluated based on completed courses measures the wrong things. The question is no longer

whether someone has attended a training course, but whether they have actually mastered the skill—and whether that mastery has been reliably documented.

Why soft skills are the most difficult category

Here lies a paradox. As technology takes over more technical work, the value of what cannot be automated continues to rise: communication, collaboration, leadership, conflict management, and delivering difficult news. These human skills—commonly referred to in practice as soft skills—also become obsolete far more slowly than technical knowledge. They are both more durable and increasingly scarce. Yet they are also the most difficult category of skills to document.

A Python certification or a driver’s license is unambiguous: you either have it or you don’t. However, “conducts a calm conversation about bad news” or “provides constructive feedback under pressure” does not appear on any diploma. These are not facts but behaviors, and behavior only reveals itself in the actual situation. Many organizations try to address this through competency-based training and detailed competency profiles. This helps define what you want to see, but a competency on a list remains just a claim. It is a promise of proficiency, not proof of it.

This brings the core of the problem into sharp focus. The skills that are most critical to an organization’s success—and that have the greatest longevity—are precisely the ones that are least visible and hardest to demonstrate. Personal development in this area often gets reduced to subjective impressions: “a pleasant colleague” or “strong communication skills.” These are fine qualities, but they are neither measurable nor transferable.

Not just an ordinary digital CV

It is tempting to view a skills passport as merely a polished CV or an expanded profile on a networking site. It is nothing of the sort, and therein lies the crucial difference. A CV is a collection of claims written by the individual; no one verifies whether a claim like “excellent communication skills” actually holds water. A skills passport worthy of the name reverses this logic: it does not show what someone *claims* they can do, but rather what they have demonstrably shown they can do.

That distinction determines its value. Badges and certificates awarded simply for attendance add little value—they confirm presence, not actual capability. Only when a passport is grounded in observable, repeated, and assessed behavior does it become more than mere window dressing. It then becomes a credible reflection of an individual’s competencies—useful for development, mobility, and deployment, both within and outside the organization.

The skills passport: from claim to evidence

At its core, a skills passport is a personal, portable record of demonstrable skills. The concept is that employees carry their competencies with them—both within and outside the organization—independent of the specific roles in which they happened to acquire them. The infrastructure for hard skills largely already exists: certificates, micro-credentials, and badges. Soft skills represent the more challenging half of the equation—and, consequently, the area where finding an effective solution yields the greatest value.

After all, a passport is only as valuable as the evidence underpinning it. A “communicative skills” checkbox without supporting evidence adds nothing more than a self-written CV. Therefore, the real question a skills-based organization must answer is not whether it can implement a passport, but how it can generate credible evidence of soft-skill behaviors—at scale, and in a way that drives development rather than merely recording it.

Practiced behavior as evidence

Such evidence is not generated in a classroom. Since Hermann Ebbinghaus described his “forgetting curve” in 1885, we have known that knowledge fades rapidly without repetition. Research into “deliberate practice” by K. Anders Ericsson (Psychological Review, 1993) shows that expertise stems from focused, repeated practice accompanied by immediate feedback. A one-day soft skills training course cannot possibly demonstrate mastery of a skill. What *can* demonstrate this is repeated, realistic practice where observable behavior is assessed.

This is where AI comes into play—not merely as a gadget, but as a solution to a measurement challenge. Through AI role-play, a person practices a conversation with an AI avatar that reacts realistically: it offers resistance, displays emotion, and continues the dialogue beyond the point where a static script would end. Because this type of practice is repeatable and scalable—and because every session provides feedback on concrete behaviors such as tone, structure, and probing questions—it generates exactly what a “skills passport” requires: a substantiated, evolving picture of what a person actually demonstrates, broken down by competency and tracked over time. AI avatar training and role-plays thus become not just a method of practice, but also the source of evidence. PractAIce is built upon this concept.

An example illustrates this concretely. Consider a team leader who struggles with delivering bad news. In an initial practice session, he rushes to the solution and leaves no room for the other person’s reaction. Two weeks and a handful of sessions later, he handles the same conversation differently: he states the message clearly, allows for a pause, and asks probing questions. That difference is neither a mere impression nor a self-assessment; it is visible in his behavior, session after session. That is precisely the building block of a skills passport: not just a checkbox, but a developmental trajectory. The difference compared to a traditional approach is fundamental. Instead of a snapshot—a one-day certificate—a continuous record of practiced behavior is created. Instead of a mere claim, there is a burden of proof. And instead of a self-contained training course, there is assurance: what has been learned is repeated, measured, and retained, rather than fading away after a week.

What this means for HR and L&D

The shift towards skills calls for a different approach to learning. Three consequences stand out:

• Focus on demonstrated competencies rather than completed courses. Make “what someone has actually shown” the unit of reporting, and link development to observable behavior instead of attendance. That is the essence of competency-based training: managing by skill, not by course.
• Design for repetition, not for a one-off event. Soft skills take root through practiced behavior spread out over time; this turns personal development into a continuous process rather than an isolated incident. Short, frequent practice sessions achieve more than a single long day of training.
• Make skills transferable and owned by the employee. A skills passport that travels with the individual boosts both development and sustainable employability.

There is a nuance here that no organization should overlook: behavioral data is for development, not for monitoring or performance management. Its value lies in making growth visible, not in holding people accountable. Those who maintain this distinction build trust rather than resistance—and trust is precisely what enables people to feel comfortable practicing their behavior.

A second caveat applies here. What you measure drives behavior; if you measure the wrong things, people will optimize for the wrong things. A skills passport must therefore be based on behaviors that truly matter, not merely on what happens to be easy to quantify. The key lies not in having *more* data, but the *right* data: behavior that reveals whether someone is genuinely conducting a conversation more effectively.

Moreover, this is not an all-or-nothing project. A sensible approach is to start small: select a few critical conversations—such as feedback sessions, delivering bad news, or sales calls—have people practice them, measure their progress, and build the passport from there. The infrastructure should follow the behavior, not the other way around.

Frequently Asked Questions

What is a skills passport?

A skills passport is a personal, transferable record of an individual’s demonstrable skills, independent of their specific job role. It shows not only the training courses a person has completed but, more importantly, the behaviors and competencies they have actually mastered—backed by evidence.

What is a skills-based organization?

A skills-based organization structures work and workforce management around skills rather than fixed job roles. Instead of job titles, the skills required to perform the work become the foundation for hiring, development, and talent deployment.

How do you make soft skills measurable?

Soft skills become measurable by repeatedly practicing and assessing observable behavior rather than inferring them from completed training courses. AI-powered role-playing makes this practice scalable and generates behavioral data for each competency, making growth visible, measurable, and transferable.

About the Author — Sven is the founder of PractAIce and a behavioral change expert. For many years, he has focused on helping organizations drive lasting behavioral change, embed new behaviors into daily practice, and make soft skills tangible and measurable.

PractAIce was built to make soft skills measurable and demonstrable. Employees practice realistic conversations with an AI Avatar, and their development is tracked and made visible for each competency. Want to see what this could look like for your organization? Discover AI Avatar Training or schedule a demo.

It is a pattern that repeats itself every Monday morning in hundreds of Dutch organizations. A group of employees returns to the workplace after a two-day soft skills training, full of new insights and good intentions. Three weeks later, there is little evidence of this in their behavior. Not because they did not want to, not because the trainer was bad. But because we have known since 1885 that the brain does not learn that way, and the type of training we organize has hardly changed despite that knowledge.

That sounds like a bold statement. It is not just an opinion. It is what cognitive psychology has repeatedly shown, and what in 2015 was once again thoroughly replicated by two researchers from the University of Amsterdam. We have known for almost a century and a half how learning sticks, and how it fades away. And yet the vast majority of soft skills training still relies on the opposite principle: one-time, intensive sessions after which people ‘just have to apply it’.

Microlearning for soft skills is all about that. Not by learning less, but by organizing it differently. Short practice sessions, spread over time, close to the moment of application. That sounds simple, but it requires tools that simply did not exist until recently.

Ebbinghaus’s forgetting curve: a discovery that organizations have hardly processed

Hermann Ebbinghaus, a German psychologist, conducted an experiment in the 1880s with himself as the only subject. He memorized lists of meaningless syllables and tested at various intervals how much he had retained. The curve that resulted from this has become one of the most well-known graphs in psychology and one of the most ignored in the world of corporate training.

The finding was shockingly simple. Within an hour of learning, Ebbinghaus had already forgotten more than half of what he had learned. Within 24 hours, about 70 percent had disappeared. What remained then leveled off at a low level that did not increase without repetition. He called this the ‘forgetting curve,’ and he demonstrated that the pattern was inevitable unless you repeated at specific times.

In 2015, the Dutch researchers Jaap Murre and Joeri Dros from the University of Amsterdam thoroughly replicated this experiment. Their study in PLoS ONE confirmed the original findings of Ebbinghaus, with modern methodology and experimental control. The curve is correct and the pattern is universal. And it says something uncomfortable about how most organizations organize their training.

Because what is a typical two-day soft skills training if not an attempt to pack as much information and practice as possible into 16 hours, only to then let those people go without structured repetition? According to science, that is approximately the least effective way to make something stick permanently.

The spacing effect: why distribution does more than intensity

In addition to the forgetting curve, there is a second finding that argues even more strongly for microlearning. The so-called “spacing effect”; the fact that the same amount of learning time, spread over multiple moments, leads to better retention than spending that same time all at once.

The most well-known evidence for this comes from a meta-analysis by Nicholas Cepeda and colleagues, who in 2006 combined the results of 184 scientific articles, together accounting for 839 independent measurements. The conclusion was clear: spaced learning leads to significantly better retention than intensive learning, almost regardless of the subject. And the effect is no small correction. For some types of material, retention doubles when the same learning time is spread over shorter blocks.

What this means for soft skills training is fundamental. A two-day training on giving feedback, no matter how well designed, can never achieve the same durability as the same 16 hours, spread over 32 half-hour sessions, distributed over a year. Science has been clear about that for almost twenty years. Practice seems to lag behind.

The question of why that is has a simple answer: until recently, there was no practical way to make that kind of distributed practice logistically possible. Booking a trainer three times a month for a half-hour session with four participants

doesn’t work. Hiring a coach for weekly short sessions with each employee is unaffordable. The idea was fine, but the infrastructure was lacking.

Why Soft Skills Are Particularly Susceptible to the Transfer Problem

Microlearning is more effective than intensive sessions for all types of subject matter, but for soft skills, the difference is particularly pronounced. This has to do with the very nature of soft skills: they are not facts to be memorized, but rather behaviors that one must be able to demonstrate in a wide variety of situations—and often at critical moments. Moreover, these situations frequently occur under pressure, or require an ad hoc response with no time for reflection.

Providing feedback to a colleague who has just made a mistake, remaining calm during a conversation that is escalating, or handling resistance without becoming defensive oneself—these are not merely items of knowledge. They are skills whose successful execution depends on self-regulation, practice, and routine. And it is precisely these skills that call for what psychologists term “deliberate practice”—a concept popularized by K. Anders Ericsson: focused, repetitive practice targeting specific aspects, accompanied by immediate feedback on one’s own actions.

What a two-day training course typically provides is a form of introduction; while useful for establishing foundational concepts, it is wholly insufficient for developing the ability to exhibit specific behaviors under pressure. Furthermore, the impact of such training is susceptible to the “illusion of competence.” By the second day of training, everyone feels more proficient than they did on day one; everyone believes they have now mastered the material. Yet it is precisely this overestimation that makes the subsequent regression—typically occurring after three weeks—all the more painful, and all the more demotivating for those wishing to attempt the learning process once again.

Consequently, soft skills—to a greater extent than other competencies—require something that traditional training methods fail to provide: a continuous rhythm of brief practice sessions, situated closely within the actual context of application, and accompanied by objective feedback. In other words: they require microlearning-based soft skills training, rather than standalone training days.

What Microlearning Is—and What It Isn’t

“Microlearning” is a term that has been widely embraced by the L&D sector in recent years—and, as a result, has also become widely diluted. For many organizations, microlearning has now come to mean short, five-minute instructional videos, an interactive quiz, or a daily “learning nudge” delivered via Teams. While not without value, this does not constitute microlearning in the sense intended by academic research.

What distinguishes microlearning for soft skills is that it is not about transferring information in small doses, but rather about practicing in small doses. A five-minute video on active listening is not practice; it is merely a shorter lecture—a transfer of knowledge. True microlearning for soft skills consists of a brief session in which an individual actively demonstrates a specific behavior, receives feedback on it, and has the opportunity to immediately try again. This represents a fundamentally different mechanism.

For organizations serious about soft skills training, this entails a re-evaluation of what “micro” actually implies. It is not about shrinking training content down into bite-sized pieces; rather, it is about increasing the frequency of practice through short, practical simulations that fit seamlessly into the daily work rhythm—whether that means five to ten minutes before a meeting, fifteen minutes on a Saturday morning, or a brief moment just before a difficult conversation.

The technological infrastructure required to make this possible has only recently become available. This also explains why microlearning for soft skills—despite the overwhelming scientific evidence supporting it—remains a relatively nascent concept in actual practice.

How AI Role-Playing Enables Microlearning for Soft Skills

This is where PractAIce comes in. The platform is built upon the very principle that cognitive psychology has advocated for over a century: short, spaced practice sessions within realistic scenarios, accompanied by immediate feedback on one’s own behavior. What was logistically impossible in traditional training is now made possible through the combination of AI avatars and customizable scenarios.

On a Tuesday morning, a manager can spend ten minutes practicing a feedback conversation with an AI avatar that responds just as a real employee would. A week later, they can do it again with a different scenario. A few days after that, they can try a variation involving greater resistance. The forgetting curve is disrupted—not by lengthy training sessions—but by brief, repeated moments occurring at the precise intervals identified by scientific research.

The results align with what research predicts. Skills that would otherwise take dozens of hours of coaching to develop can now be systematically built into a regular work schedule. An employee who devotes fifteen minutes each week to a role-playing exercise gains more practice in a single year than most colleagues do in their entire careers. And most importantly: this practice aligns with how the brain learns, rather than with how schedulers plan.

For L&D professionals, PractAIce adds a dimension that traditional microlearning lacks: measurability at the behavioral level. For every practice session, the platform generates data on an individual’s performance regarding specific competencies—for instance, the specificity of their feedback, how they handled emotions, or how they structured the conversation. This data, collected over weeks and months, reveals a team’s developmental trajectory in a way that was previously impossible.

Frequently Asked Questions about Microlearning and Soft Skills

What exactly is microlearning?

Microlearning is a form of learning in which short, focused learning moments are distributed over time, rather than concentrated into long sessions. For soft skills, this translates to short practice sessions—lasting five to fifteen minutes—conducted repeatedly, featuring varied scenarios and immediate feedback. It differs fundamentally from “short instructional videos”: with microlearning, the emphasis lies on active practice, not passive consumption.

Can you really develop soft skills in short sessions?

Yes—and scientific research even suggests that doing so is more effective than using long sessions. The crucial element is not the duration of a single session, but rather its frequency and how it is distributed over time. Ten fifteen-minute sessions spread out over a month consistently yield more lasting change than two consecutive days of training. This holds even truer for soft skills than for acquiring factual knowledge, as behavioral change relies even more heavily on repeated application.

How often should microlearning take place to be effective?

Research into the “spacing effect” suggests that the optimal frequency depends on how long the material needs to be retained. For skills you wish to maintain continuous mastery of, a frequency of one to three times per week works well. What matters more than the exact frequency, however, is consistency. A weekly practice session that is sustained over time is more effective than a daily session that is abandoned after just two weeks.

Does microlearning replace traditional soft skills training?

Not entirely, but it does shift where the primary value lies. Traditional training remains valuable for initially introducing a topic, establishing a shared vocabulary within a team, and discussing complex situations that require human nuance. Microlearning fills a gap that traditional training could never quite bridge: the structured practice required to translate knowledge into actual behavior. The combination is stronger than either component alone.

Does microlearning work for leadership and management as well?

That is precisely where it works particularly well. Leadership skills—switching between styles, handling resistance, conducting difficult conversations—are exactly the type of skills that develop most effectively through short practice sessions. A manager who spends ten minutes each week practicing a challenging conversation with an AI avatar will, over the course of a year, build a behavioral repertoire that is simply unattainable through traditional training.

In Conclusion

The science surrounding how people learn and forget is not a new discovery. Ebbinghaus conducted his experiments at a time when people were still cooking on coal stoves. These principles have been repeatedly confirmed—most recently by Dutch researchers. The spacing effect is one of the most robust findings in cognitive psychology. Anyone who takes soft skills training seriously can no longer afford to ignore these insights.

Until recently, what was missing was the infrastructure to put these insights into practice. Organizing short practice sessions—spaced out over time, featuring immediate feedback and realistic scenarios—was simply not scalable within most organizations. This is precisely what AI role-playing changes. It is not merely a futuristic pipe dream, but a concrete, readily available method for aligning soft skills training with the way the human brain actually learns.

Wondering how microlearning for soft skills training would work in your organization? A demo of PractAIce takes just fifteen minutes to show you how a short practice session unfolds and the type of developmental data the platform accumulates over time.

If you were to request the personnel administration records of a random Dutch organization, you would receive a list of names and job titles. Senior Consultant. Operations Team Lead. Junior Account Manager. It is so commonplace that we barely stop to think about it, but job titles have been the primary organizing principle of work for the past hundred years. Who you are within an organization is largely determined by what is printed on your business card.

And that exact principle is now under pressure from a generation that fundamentally views it differently. Gen Z, the generation born roughly between 1995 and 2010, is the first generation in the workplace that no longer sees the vertical ladder as self-evident. And that has much deeper consequences for competency-based work than most organizations realize.

This is not another “Gen Z is different” argument. It is an observation about what is actually shifting underneath that discussion: from a work model centered around roles and titles to a work model centered around what people can do, what they want, and what they want to develop. For L&D professionals, this is not some distant future. It is happening now.

What the data shows and why the bookshelf is starting to wobble

It is tempting to dismiss Gen Z’s behavioral shift as generational complaining. The problem is that the data is becoming increasingly consistent. Only 6% of Gen Z identifies reaching a senior leadership position as their primary career goal, according to Deloitte’s Global 2025 Gen Z and Millennial Survey. At the same time, learning and development consistently ranks in the top three reasons they choose an employer. They want to grow, just not necessarily upward (vertically).

What this reveals is something more fundamental than a shift in ambition. Gen Z has implicitly realized that vertical promotion is becoming less predictive of what work actually is. The role “team leader” means something entirely different in ten different organizations. A senior title says little about what someone can actually do in 2026, considering how quickly work itself changes. And a career plan like “in five years I want to be X” rarely matches the reality of what someone is actually doing five years later.

The World Economic Forum confirmed this in its Future of Jobs Report 2025: 39% of employees’ current skills will be transformed or outdated by 2030. In essence, that means a complete renewal of workforce skills every decade. A job title is not flexible enough to keep pace with that speed, but a competency profile is.

The quiet shift toward competency-based work

What Gen Z is intuitively doing is now being structurally organized by forward-thinking companies. It is called the “skills-based organization,” or in Dutch terms: competency-based work. The idea is simple in theory, complex in execution: stop organizing around roles and start organizing around competencies. Instead of an org chart where people are locked into fixed roles, organizations become dynamic networks where people are deployed based on what they can do, and where development focuses on the competencies they want to build.

Deloitte’s own research calls this “a portfolio of ways to organize work, enabling greater agility and more meaningful packages of work.” It sounds abstract, but the impact is concrete: internal mobility moves faster, talent is used more effectively, and employees stay longer because growth is no longer tied to an empty chair above them. The World Economic Forum even stated that skills-based hiring works for jobs that do not yet exist, because the focus shifts from what someone was to what someone can become.

For organizations making this transition, a paradox emerges. On one hand, it delivers strategic advantages leadership teams love: agility, retention, and better use of talent. On the other hand, it requires something most organizations are not prepared for: infrastructure that makes competencies visible, trainable, and measurable. Without that infrastructure, “skills-based” remains a nice phrase on a strategy slide rather than a reality on the work floor.

The problem L&D now has to face

This is where things become uncomfortable for L&D professionals. Most learning and development programs in Dutch organizations are still organized around roles, not competencies. A training offer is called “Leadership for New Managers,” not “Giving feedback to defensive team members.” A training catalog is usually a list of courses, not a map of interconnected competencies.

The result is that training remains an event tied to a role transition, rather than an ongoing process connected to the competencies someone wants to develop at a specific moment. For a Gen Z employee who wants to grow without necessarily becoming a manager, that offering feels irrelevant. For an organization aiming to work competency-first, it lacks the actual building blocks.

What is missing is what researchers from the World Economic Forum call a “Global Skills Taxonomy” for the individual organization: a shared overview of which competencies matter, how they connect, how they are developed, and how growth can be observed. Not as a bureaucratic instrument, but as an operational framework for recruitment, development, and internal mobility. Many organizations start building this but run into a deeply practical issue: how do you make soft skills competencies visible at all?

Soft skills competencies: making visible what has always been intangible

This is the heart of the challenge. For hard competencies, such as mastering a programming language, writing an Excel formula, or giving a presentation in English, there are ways to demonstrate proficiency. A test, a project, a certificate. But for the competencies growing fastest in Future of Jobs rankings — think effective communication, handling resistance, situational leadership, or conflict management — reliable measurement tools have long been missing.

That is not accidental. Soft skills competencies are contextual. Someone can be brilliant at giving feedback to a peer and completely freeze in a conversation with a defensive executive. Someone can remain calm in a meeting and escalate emotionally in a one-on-one conversation. A single snapshot says very little; you need to see how someone behaves across different situations to reliably assess competency.

Until recently, this meant: 360-degree feedback, manager observations, peer review. All valuable, all labor-intensive, and none scalable enough to support competency-based work across an entire organization. It is exactly this gap where AI training and AI roleplay introduce something fundamentally new. Not as a gimmick, but as infrastructure for the competency-based work model Gen Z already expects.

How AI roleplay makes competencies visible and trainable

PractAIce connects directly to this challenge. The platform allows employees to go through conversational scenarios with an AI avatar that responds the way a real colleague would. For example in feedback conversations, conflict situations, or negotiation moments. What this changes is not the content of training, but its structure. Instead of isolated role-based training sessions, organizations create an ongoing practice environment centered around competencies.

An employee who wants to improve feedback skills can practice across a range of scenarios that reveal the complexity of that competency. Someone focused on situational leadership can move through scenarios where different leadership approaches are appropriate. The scenario aligns with the employee’s development goals, not the schedule of an external training provider.

And perhaps just as importantly, PractAIce generates behavioral-level data from every practice session. How specific was the feedback? Was enough space given to the other person? How was defensiveness handled? That data, collected across weeks and teams, creates something that previously did not exist: a reliable measurement system for soft skills competencies that does not depend on self-reporting or a manager’s temporary impression. For a competency-based organization, that is not a luxury — it is a requirement.

It also matches the learning rhythm Gen Z expects. Short practice sessions, completed when convenient, with immediate feedback. Not a yearly training event, but continuous micro-moments in which competencies gradually develop. That is not a “Gen Z adjustment”; it is how the brain actually learns. Gen Z just happens to be the first generation demanding it explicitly.

What L&D can do now without turning everything upside down

The transition to competency-based work does not need to be a big-bang transformation. What helps is starting with three concrete shifts, each valuable on its own and together capable of moving something much deeper.

The first shift is linguistic: stop talking in roles and start talking in competencies. In development conversations, performance reviews, and job descriptions. Not “I want to become senior,” but “I want to improve my ability to handle difficult conversations.” At first that language feels forced, but within months it changes how people think about growth.

The second shift is structural: create explicit learning paths for the most critical soft skills competencies, separate from roles. A learning path for giving feedback that applies equally to a team lead, an individual contributor, and a director. Tools like PractAIce make this scalable by adapting scenarios to the user’s level and context.

The third shift is measurable: introduce competency data into development conversations, not as an evaluation tool, but as a growth tool. Not “how are you scoring?” but “can you see where you are improving and where refinement is possible?” The shift sounds small, but it fundamentally changes how employees view their own development. And it aligns perfectly with Gen Z’s existing relationship with data about sleep, fitness, and personal habits.

Frequently asked questions about competency-based work

What is the difference between competency-based work and role-based work?

Role-based work organizes people around positions with fixed responsibilities and tasks. Competency-based work organizes people around their skills, mindsets, and development potential. The difference is not just terminology but emphasis: in a competency-based organization, someone can contribute in multiple ways based on what they can do, rather than what their title allows. For internal mobility, talent retention, and organizational agility, that is a fundamentally different way of working.

Does competency-based work also function in traditional industries, or only in tech companies?

The principles work everywhere, but the implementation differs. In highly regulated sectors such as healthcare, education, and government, job titles remain formally important, but competency-based work can coexist in how development, mobility, and evaluation are organized. It does not require a complete restructuring; it can function as a layer on top of the existing structure.

How do you reliably measure soft skills competencies?

The weakest method is self-reporting; asking someone “are you good at giving feedback?” produces little useful data. Stronger methods include 360-degree feedback and structured observation by managers, but those are labor-intensive and highly dependent on specific moments. AI roleplay adds a third layer: standardized behavioral observation in scenarios that genuinely challenge the competency, measured across multiple practice sessions. That creates a richer picture than any single method alone.

How does AI training fit into a competency-based organization?

AI training works best as infrastructure for competency development, not as a standalone solution. PractAIce makes it possible to build practice pathways for specific soft skills competencies — such as effective communication, handling resistance, or situational leadership — that demonstrate development over time. Training then becomes part of the rhythm of work itself, not a separate event category.

In conclusion

The end of the job title sounds more dramatic than it actually is. Job titles will continue to exist; they are practical, legally embedded, and culturally ingrained. What is changing is their position as the primary organizing principle. For the generation now entering the workforce, and increasingly for the organizations trying to attract them, the key question is no longer what someone is, but what someone can do and wants to develop.

For L&D professionals, this is an invitation to rethink the foundations of the profession. Not to tear everything down, but to build the infrastructure that truly enables competency-based work: a shared language of competencies, learning pathways independent of roles, and data that makes behavioral development visible.

Would you like to explore how PractAIce could fit into your organization as infrastructure for competency-based work? A demo shows in fifteen minutes how a practice session works and what kind of competency data the platform builds over time.

Until around 1820, almost nobody learned anything through a course. Anyone who wanted to make shoes became an apprentice to a shoemaker. Anyone who wanted to work with metal worked alongside a blacksmith. Anyone who wanted to learn writing did not mainly read about writing, but sat next to someone who actually did it. The system was called “master and apprentice,” and it worked so well that it was independently invented in nearly every civilization. For centuries, it was the most natural way for people to acquire craftsmanship.

That system has largely disappeared from the modern working world. Not because it worked worse than what replaced it, but because it was not scalable to the new industrial reality. A master could guide one or two apprentices, not thousands of employees. We compensated for that with courses, books, education, and later with training programs. And we silently accepted that, for modern and complex skills, there would never again be a master standing beside you while you did the work.

That is exactly what is changing now. AI coaching is not a break from the way we learn; in a sense, it is a return to it — only on a scale that would have been unimaginable in 1820. And for soft skills training, that means something fundamental.

What we discovered when we took another look at learning in practice

In 1989, three scientists investigated why people in the past often learned faster and better through practice. Their conclusion was simple: people learn far more effectively when someone is beside them to demonstrate, observe, correct, and help in the moment itself. Not just explanation beforehand, but guidance during the actual doing. They later gave this way of learning the name “cognitive apprenticeship.”

What they uncovered was that a good master did six things that rarely come together in modern instruction: demonstrating (modeling), guiding and correcting during execution (coaching), providing temporary support (scaffolding), encouraging the learner to explain their thinking (articulation), stimulating reflection, and ultimately creating space for independent experimentation. According to Collins and his colleagues, it is precisely the combination of these six elements that makes the difference between superficial learning and lasting learning.

What fascinated them most was why this model was so rarely applied in practice. The answer lay in scale: a teacher cannot treat a class of 30 students the way a shoemaker treats a single apprentice. As a result, the form of education lagged behind what we already knew to be more effective. For soft skills, the gap between what we knew and what we actually did was even greater: how do you simulate a master standing beside you during a difficult conversation when that master also has their own work to do in the evening?

Situated learning: why learning depends on context

Around the same time, John Seely Brown, Allan Collins, and Paul Duguid developed a related idea that became known as “situated cognition.” Their argument — which sounded radical in 1989 but is now considered almost mainstream — was that knowledge and context are inseparably connected. Learning something in a classroom and applying it in the workplace are not two phases of the same process; they are two different learning processes.

Their famous formulation stated that “situations partially produce knowledge through activity.” In other words: what you learn is partly shaped by where you learn it. Someone who learns conflict management through a role-play exercise with a colleague in a training room learns something fundamentally different from someone who learns it during a real conversation with a frustrated customer. Not because the content is different, but because the context causes the brain to encode it differently.

For soft skills, this has an uncomfortable implication. The skills you practice during a training day are difficult to transfer to the workplace because the workplace is a different context from the training room. That does not mean the training was of poor quality, but rather reflects a fundamental characteristic of how learning works. It explains the transfer problem that L&D has struggled with for decades: not because trainings are ineffective, but because they miss one crucial element — proximity to the real context of execution.

It is exactly what that old way of learning did have. People did not practice in a separate space, but in the same context in which the real work took place. The result was craftsmanship that lasted, because learning and doing were never separated from one another.

What AI coaching now makes possible

This is where two developments come together that are each interesting on their own, but together create something entirely new. AI is capable of simulating realistic conversation scenarios in which behavior — not just knowledge — can be practiced. And that same AI is available at the moments when a human mentor never could be, for example at 7:30 in the morning before a difficult meeting or at 10 o’clock at night after a day in which something went wrong.

Voor het eerst sinds we praktijkleren hebben vervangen door klassikaal leren, kunnen we de principes achter die oude leervorm opnieuw toepassen, maar nu op een schaal die vroeger onmogelijk was. AI-coaching maakt dat mogelijk. Het kan tijdens de oefening bijsturen. Het kan ondersteuning afbouwen naarmate iemand vaardiger wordt. Het kan de leerling laten verwoorden wat ze deed en waarom. En het kan dit doen voor honderden medewerkers tegelijk, zonder dat de kwaliteit per oefening daalt.

Importantly, AI coaching does not replace a human coach or mentor. Just as the master in 1820 was not replaced by a textbook, AI does not replace the human nuance, life experience, and wisdom that a good coach provides. What AI does offer, however, is something that had largely disappeared from the modern workplace at scale: direct, contextual practice with immediate feedback, available at the moment it matters most.

Why this is the biggest change in soft skills training in a hundred years

For hard skills, we have had effective learning structures for quite some time. You learn programming by writing code and seeing it work or fail. You learn English grammar by writing sentences that a corrector reviews. The feedback loop is fast, and the result is immediately measurable.

For soft skills training, this was fundamentally different until recently. Someone could spend hundreds of hours in training about giving feedback and only discover in practice that they still could not do it effectively. The feedback loop was often delayed rather than immediate, sometimes taking days or weeks instead of minutes. That is one of the reasons why soft skills, despite decades of training, have not structurally improved in most organizations.

AI training with realistic avatars changes that dynamic completely. A manager can practice a feedback conversation twenty times before having it in real life. They can experiment with openings, wording, and different ways of handling resistance. And after every attempt, they receive immediate, behavior-specific feedback. This is not an incremental improvement over traditional training; it is a fundamentally different learning process, much closer to the way the brain actually develops skills.
PractAIce is built around exactly this principle. The AI avatar is not just a conversation partner; it is a partner that operationalizes the six elements of cognitive apprenticeship. The platform demonstrates what effective behavior looks like, provides scaffolding for beginners, offers explicit reflection after every exercise, and gradually reduces support as the user gains mastery. That is not accidental. It is built on four decades of learning science.

What this means for L&D professionals

The practical implication is that the role of L&D within organizations is shifting. No longer merely the organizer of training sessions, but the architect of learning infrastructure. No longer simply the booker of trainers, but the builder of practice pathways. No longer dependent on one-time interventions, but responsible for continuous development in the flow of work.

That requires different questions at the start of a learning journey. Not “which training fits this role transition?”, but “which competencies do we want to develop, and how do we build a practice structure that operationalizes the six elements of cognitive apprenticeship?” That may sound academic, but in practice it is surprisingly practical. Especially with platforms that can do the kind of work a human mentor simply cannot sustain for hundreds of employees at the same time.

The beauty of this shift is that it does not require replacing everything that already exists. Traditional training remains valuable for introducing knowledge and creating a shared language. Coaching remains essential for human nuance. What AI coaching adds is the practice infrastructure in between — the part where most behavioral change actually takes place, and which until now has simply been missing in most organizations.

Frequently asked questions about AI coaching and the future of learning

What is the difference between AI coaching and e-learning?

E-learning is, at its core, the transfer of information in digital form: videos, modules, quizzes. The brain consumes the material, but does not practice behavior. AI coaching is fundamentally different: it is interactive, scenario-based, and focused on behavioral change. Not learning about giving feedback, but actually giving feedback and receiving immediate responses to it. That difference explains why AI coaching has a much stronger impact on soft skills than traditional e-learning ever could.

Does AI coaching replace human coaches?

No. They serve different functions. A human coach provides context, life experience, and intuition about what is happening beneath the surface — things AI cannot replicate. AI coaching provides scalable practice opportunities with immediate feedback on specific behaviors, something a human coach could never sustain for hundreds of employees. The combination is stronger than either on its own: human coaching for strategic conversations, AI training for continuous practice.

How does the brain actually learn soft skills?

Not by reading or listening about them, but through application. Soft skills are behavioral patterns, and behavioral patterns develop through repeated execution combined with feedback. Cognitive psychology has been clear about this for more than a century. What is new is that we now have an infrastructure that makes this repetition possible without depending entirely on a human coach — and that opens possibilities that simply never existed for most organizations.

Does AI coaching fit every role, or mainly leadership?

It works especially well for leadership, because leadership skills depend heavily on switching between styles under pressure. But its application is much broader: customer-facing roles, sales, healthcare, and any function where conversational ability shapes performance. What determines its relevance is not the seniority of the role, but whether it involves soft skill competencies that can only be developed through practice.

In conclusion

The history of learning is long, and most shifts within it have been relatively small. What we are experiencing now is bigger than a passing trend. For the first time in two centuries, we can once again make the form of learning that has always been the most effective — direct practice with a master standing beside you — available at the scale required in the modern workplace.

This is not a solution to everything. It requires thoughtful implementation, realistic expectations, and a clear understanding of what AI can and cannot do. But for the first time, it opens the possibility of approaching soft skills training in the way learning science has long prescribed: not as an event, but as a practice; not as knowledge transfer, but as behavioral development; not as an annual workshop, but as a continuous learning process.

If you would like to explore how AI coaching through PractAIce could work within your organization, a demo can show in fifteen minutes how a conversation with an AI avatar works — and provide a far more concrete impression than any description ever could.