Across major workforce studies, human capabilities such as adaptability, creative thinking, collaboration, and leadership are becoming increasingly important. As AI increasingly automates parts of routine and codifiable work, it’s the distinctly human capabilities that determine whether a team can actually use its new tools well. McKinsey’s newest research, The State of Organizations 2026, drawn from more than 10,000 senior executives across 15 countries, names three “tectonic forces” now reshaping organizations: technology infusion, economic and geopolitical disruption, and changing workforce expectations, demographics, and work models (McKinsey & Company, 2026).
Boston Consulting Group’s 2026 AI Radar reveals a related confidence gap: 62% of CEOs say they’re confident their AI investments will eventually pay off, compared with just 48% of non-tech executives outside the C-suite. BCG suggests that executives closer to day-to-day AI work may have a more realistic view of implementation challenges (BCG, 2026). BCG’s own conclusion is blunt: “success depends on people and change management, not just technology” (BCG, 2026). Deloitte’s 2026 Global Human Capital Trends report puts a number on that same gap: 85% say it’s critical to build the organization’s and workforce’s ability to adapt at the speed today’s world requires, but only 7% say they’re leading in this area (Deloitte, 2026). The World Economic Forum, meanwhile, estimates that employers expect 39% of workers’ core skills to change by 2030, with creative thinking and resilience/flexibility/agility among the fastest-rising in importance (World Economic Forum, 2025).
So the trend lines agree that soft skills matter, more and more urgently. Far fewer organizations have a real plan for developing them.
That’s the gap our white paper, Soft Skills as Strategic Competencies in the AI Era, digs into: organizations still devote much of their formal learning effort to technical and compliance content, while the relational, reflective learning that actually builds soft skills gets a single workshop and a pat on the back. Reading about empathy doesn’t make you more empathetic. Attending one seminar on giving feedback doesn’t make you better at it. So what does the science actually say works?
Why knowing isn’t the same as doing
Hard skills – a new CRM, a spreadsheet formula, a compliance procedure – can often be transferred through instruction: read it, watch it, apply it. With soft skills, this alone is not enough, and the transfer problem becomes especially visible. They are personal and interpersonal capabilities whose expression depends heavily on social interaction and situational context (Cimatti, 2016). You don’t learn to give better feedback by hearing a definition of good feedback; you learn it by giving feedback, in a real situation, and finding out – from another person – what landed and what didn’t.
This tracks a distinction that’s well established in training research: acquiring knowledge and transferring it into actual job behavior are not the same thing. Classic work by Baldwin and Ford (1988), and the large-scale meta-analytic follow-up by Blume, Ford, Baldwin, and Huang (2010), show that whether training generalizes into real workplace behavior depends heavily on factors beyond the training content itself – how the learning experience is designed, whether people get the chance to actually apply the new skill, and the support and cues present in the work environment. Content delivery alone – reading, watching, a single workshop – is unlikely to be enough when the desired outcome is behavioral transfer into complex interpersonal situations. Digital learning platforms have made content dramatically more accessible; the transfer research suggests that access to content alone does not solve the core transfer problem in soft-skills development.
Blume and colleagues’ analysis of 89 studies makes this especially clear for a distinction training researchers call “open” versus “closed” skills. Closed skills, like operating a piece of software, have one correct way to perform them; open skills, like leadership or interpersonal communication, require trainees to adapt and exercise judgment in applying them. Motivation and a supportive work environment predicted transfer far more strongly for open skills (r = .19 and .26, respectively) than for closed skills (r = .11 and .04) – meaning the things that make training “stick” matter more, not less, for exactly the kind of skills this article is about.
Recent workplace-learning surveys describe a similar gap from a different angle. TalentLMS’s 2026 L&D Report finds that 86% say they pick up skills or knowledge by figuring things out on the job, yet close to a third describe the training they actually receive as too theoretical (TalentLMS, 2026) – survey data that captures the current state well, even if, as an industry survey rather than controlled research, it can’t establish causation the way the transfer literature can. The same report shows the share of organizations offering soft-skills training rising from 38% in 2024 to 47% in 2025, after several years of decline (TalentLMS, 2026).
Reflection is where it starts, but it has a boundary
There’s good reason to start with the individual. Research on leadership self-awareness has repeatedly linked greater awareness of one’s own behavior – including alignment between self- and other-perceptions – with leadership-related outcomes (Atwater & Yammarino, 1992; Ashley & Reiter-Palmon, 2012). Tekleab, Sims, Yun, Tesluk, and Cox (2008) found that when leaders’ self-ratings agreed with follower ratings, transformational-leadership self-awareness was linked to higher leader effectiveness and follower satisfaction, while empowering-leadership self-awareness was linked instead to followers developing stronger self-leadership. A longitudinal, mixed-method evaluation of workplace self-awareness training by Sutton, Williams, and Allinson (2015) found that structured interventions can produce a measurable shift in self-awareness over a six-week period.
Personal reflection also has an inherent limitation, though: it operates within the individual’s existing interpretations and blind spots. Feedback and dialogue can introduce information and perspectives that solitary reflection cannot generate on its own. That’s where the evidence gets substantially stronger – and where it points toward other people, in the room, talking about the actual work.
The stronger predictor: shared, context-bound reflection
Team reflexivity – the habit of a team pausing together to discuss its goals, processes, and results, and adjusting accordingly – is one of the more thoroughly studied constructs in organizational psychology. In an early study of 98 primary health care teams, Schippers, West, and Dawson (2015) found that team reflexivity was associated with greater innovation, particularly in more complex and demanding work environments. A recent meta-analysis by Leblanc, Harvey, and Rousseau (2024), synthesizing decades of subsequent research, finds an overall positive relationship between team reflexivity and team performance – while also showing that the strength of that relationship depends on specific team and task conditions. Related meta-analytic work on structured debriefing – teams deliberately stopping to review what just happened – found that well-implemented debriefs improve subsequent performance by roughly 20–25%, an effect found across a wide range of simulated and real-world settings (Tannenbaum & Cerasoli, 2013).
Two mechanisms help explain why this kind of shared, in-context reflection tends to matter.
First, it’s context-bound. The reflection happens on the actual problem the team just faced, not a hypothetical case study. The training-transfer research cited above (Baldwin & Ford, 1988; Blume et al., 2010) helps explain why this matters: whether learning generalizes into job behavior depends not only on what’s taught, but on the opportunity to apply it and the cues present in the real work environment. A team that just struggled to align on a decision, and immediately reflects together on why, is applying – and therefore learning – exactly the skill it needs, in the environment it needs to use it in. That’s a fundamentally different proposition from a generic communication-skills course delivered months earlier to a room of strangers.
Second, it’s social – which means it comes with feedback. Feedback is one of the most extensively studied influences on learning. In educational research, Hattie and Timperley’s foundational synthesis (2007) and the large-scale meta-analytic update by Wisniewski, Zierer, and Hattie (2020) – covering 435 studies, 994 effect sizes, and more than 61,000 participants – both find a meaningful overall effect of feedback on learning outcomes, though the size of that effect depends strongly on the type and informational quality of the feedback given. Shared reflection also creates opportunities for feedback: it’s the moment where you find out how your communication actually landed, or where a colleague names a blind spot you couldn’t see yourself.
Psychological safety supports and enables this kind of effective reflexivity: the shared sense that it’s safe to take interpersonal risks, ask questions, and admit mistakes. In its own account of Project Aristotle – a multi-year study of more than 180 teams – Google describes psychological safety as the most important of the team dynamics it examined (Google re:Work, n.d.; see also Duhigg, 2016). Edmondson’s original research (1999) established the construct and its link to team learning behavior. Shared reflection is far more likely to produce honest, useful feedback in a climate where people aren’t afraid to be candid. When psychological safety is low, candid reflection becomes harder, and discussions are more likely to stay superficial or turn defensive.
Putting it together: what actually builds soft skills
Read as a chain rather than a list of citations, the research says something fairly specific:
- Soft skills require behavioral transfer, not just knowledge acquisition (Baldwin & Ford, 1988; Blume et al., 2010) – so content delivery alone was never going to be enough.
- Individual reflection helps people notice their own reactions and assumptions, building the kind of self-awareness linked to better leadership and follower outcomes (Atwater & Yammarino, 1992; Ashley & Reiter-Palmon, 2012; Tekleab et al., 2008; Sutton et al., 2015) – but it’s bounded by the individual’s own blind spots.
- That boundary loosens when reflection becomes social. Shared, context-bound reflection is associated with better team performance and innovation (Schippers et al., 2015; Leblanc et al., 2024), and structured debriefing produces some of the largest, most consistent performance effects in this literature (Tannenbaum & Cerasoli, 2013).
- This is supported and enabled by psychological safety (Edmondson, 1999; Google re:Work, n.d.).
- Insight needs to convert into a specific, planned action, or it tends to evaporate. Decades of research on implementation intentions – simple, concrete “if–then” plans specifying when, where, and how a new behavior will be enacted – show that this step reliably increases the odds that intention turns into actual behavior (Gollwitzer, 1999; Gollwitzer & Sheeran, 2006).
At Academy of Brain, this is the logic behind our Reflective Cycle Framework: a repeatable loop of learning spark → personal reflection → collective meaning-making → collaborative action planning, run in minutes, inside real teams, on real work – not as an annual off-site, but as a habit. To be clear about what kind of claim this is: the Reflective Cycle itself has not been independently tested or validated as an intervention. It’s a research-informed synthesis of five separate, well-established research lines, where each step maps onto a specific piece of the evidence chain above – not a study result in its own right. A short learning spark provides a focused, common input for reflection; microlearning research suggests that brief, focused content can support retention and reduce cognitive load (Mostrady et al., 2024), but the spark is delivery architecture, not the mechanism that changes behavior. The mechanism is what happens after it: reflection, made social, inside safety, converted into a plan.
Deloitte’s 2026 Human Capital Trends report arrives at a similar structural conclusion from the industry side: adaptation and learning should increasingly be embedded into the flow of work through hands-on practice, real-time support, and continuous feedback loops rather than treated primarily as separate training events (Deloitte, 2026) – a close match for the bet the Reflective Cycle makes.
The takeaway
If you’re responsible for building soft skills in your organization, the evidence points away from one-off workshops and toward small, structured, recurring moments where teams reflect together on their actual work and give each other real feedback, inside a climate where that’s safe to do. Individual reflection is where it starts. Shared, context-bound reflection – backed into a concrete next step – is where it compounds.
Sources
Ashley, G. C., & Reiter-Palmon, R. (2012). Self-awareness and the evolution of leaders: The need for a better measure of self-awareness. Journal of Behavioral and Applied Management, 14(1), 2–17.
Atwater, L. E., & Yammarino, F. J. (1992). Does self–other agreement on leadership perceptions moderate the validity of leadership and performance predictions? Personnel Psychology, 45(1), 141–164.
Baldwin, T. T., & Ford, J. K. (1988). Transfer of training: A review and directions for future research. Personnel Psychology, 41(1), 63–105.
Blume, B. D., Ford, J. K., Baldwin, T. T., & Huang, J. L. (2010). Transfer of training: A meta-analytic review. Journal of Management, 36(4), 1065–1105.
Boston Consulting Group. (2026). AI Radar 2026: As AI investments surge, CEOs take the lead. BCG
Cimatti, B. (2016). Definition, development, assessment of soft skills and their role for the quality of organizations and enterprises. International Journal for Quality Research, 10(1), 97–130.
Deloitte. (2026). 2026 Global Human Capital Trends: Winning organizations will build the human advantage. Deloitte
Duhigg, C. (2016). What Google learned from its quest to build the perfect team. The New York Times Magazine.
Edmondson, A. C. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44(2), 350–383. Read on HBS
Gollwitzer, P. M. (1999). Implementation intentions: Strong effects of simple plans. American Psychologist, 54(7), 493–503.
Gollwitzer, P. M., & Sheeran, P. (2006). Implementation intentions and goal achievement: A meta-analysis of effects and processes. Advances in Experimental Social Psychology, 38, 69–119.
Google re:Work. (n.d.). Understand team effectiveness. rework.withgoogle.com
Hattie, J., & Timperley, H. (2007). The power of feedback. Review of Educational Research, 77(1), 81–112. Read the paper
Leblanc, P.-M., Harvey, J.-F., & Rousseau, V. (2024). A meta-analysis of team reflexivity: Antecedents, outcomes, and boundary conditions. Human Resource Management Review, 34(4), 101042. https://doi.org/10.1016/j.hrmr.2024.101042
McKinsey & Company. (2026). The state of organizations 2026: Three tectonic forces that are reshaping organizations. McKinsey
Mostrady, A., Sanchez-Lopez, E., & Gonzalez-Sanchez, A. (2024). Microlearning and its effectiveness in modern education: A mini review. Acta Pedagogia Asiana, 4(1), 33–42.
Schippers, M. C., West, M. A., & Dawson, J. F. (2015). Team reflexivity and innovation: The moderating role of team context. Journal of Management, 41(3), 769–788. https://doi.org/10.1177/0149206312441210
Sutton, A., Williams, H. M., & Allinson, C. W. (2015). A longitudinal, mixed-method evaluation of self-awareness training in the workplace. European Journal of Training and Development, 39(7), 610–627.
Tannenbaum, S. I., & Cerasoli, C. P. (2013). Do team and individual debriefs enhance performance? A meta-analysis. Human Factors, 55(1), 231–245. PubMed
TalentLMS. (2026). The TalentLMS 2026 L&D report: The state of workplace learning. TalentLMS
Tekleab, A. G., Sims, H. P., Yun, S., Tesluk, P. E., & Cox, J. (2008). Are we on the same page? Effects of self-awareness of empowering and transformational leadership. Journal of Leadership & Organizational Studies, 14(3), 185–201.
Wisniewski, B., Zierer, K., & Hattie, J. (2020). The power of feedback revisited: A meta-analysis of educational feedback research. Frontiers in Psychology, 10, 3087. Frontiers
World Economic Forum. (2025). The future of jobs report 2025.
Ojanen, V. (2026). Soft Skills as Strategic Competencies in the AI Era. Academy of Brain White Paper.
Subject matter by Ville Ojanen
(Lead Psychologist at Academy of Brain)
Ville Ojanen is a psychologist with a doctorate in cognitive neuroscience. His passion is understanding and influencing the human behavior. Ville combines an exceptionally broad and in-depth understanding of behavioral and neurosciences with practical experience in developing individuals, teams and communities.
He has worked as a clinical neuropsychologist, as an HR professional in Nokia and for fifteen years as an organizational psychologist, coach and keynote speaker. Ville is a serial entrepreneur and the co-founder of Academy of Brain Ltd. He has served in the behavioral science advisor team of the Finnish Prime minister´s office since 2020, currently as a steering group member.
