On this episode of Meaningful Work Matters, Andrew speaks with Llewellyn van Zyl, a positive organizational psychologist and data scientist who is reshaping how we think about employee well-being.
As a professor at North-West University and Chief Solutions Architect at Psynalytics, van Zyl combines deep expertise in positive psychology with hands-on experience in analytics and machine learning.
In this conversation, he explains why traditional top-down models of well-being often fall short, and introduces a bottom-up, person-centered approach that treats every individual as unique. He also explores how artificial intelligence might help scale these insights, what risks and ethical concerns come with that, and what it all means for the future of work.
Why Top-Down Models Fall Short
For decades, much of positive psychology has relied on “top-down” models of well-being, such as the PERMA framework. These approaches assume that experts can define the components of well-being, design measurement tools, and then apply them across contexts.
While useful for prediction and creating shared language, van Zyl argues that these models break down in practice.
The problem is that what counts as well-being is not universal. A framework developed in the United States may not hold in Saudi Arabia, or even across subgroups within the same country. In one study, half of the psychological strengths identified by LGBTQ+ participants did not appear in the VIA Strengths model. Pride, for example, emerged as a critical strength, even though psychology often labels it as a vice.
Context matters, and averages fail to capture the lived realities of individuals.
Top-down models also treat well-being as static, impose narrow categories, and depend on self-report measures that often mask what people are actually experiencing. Van Zyl points out that someone may show all the physiological signs of burnout while still rating themselves as “a little stressed” on a survey.
A Bottom-Up Approach to Well-Being
Van Zyl offers an alternative: a bottom-up, person-centered perspective.
Instead of imposing categories from above, this approach starts with the individual and builds outward. He describes eight principles that make up this way of thinking:
Every person is unique. Each individual is a case study of one, shaped by their own history, drivers, and experiences.
Context matters. External conditions directly influence well-being.
We are embedded in systems. Families, workplaces, and policies both shape and are shaped by individuals.
Well-being is a process, not a snapshot. It unfolds over time and fluctuates in response to experiences.
We need multiple perspectives. Stories, objective indicators, physiological data, and even traditional surveys all contribute to a fuller picture.
It is co-created. Insights emerge through dialogue between individual and practitioner, not imposed by one on the other.
Meaning is personal. The same factor can enhance one person’s well-being while detracting from another’s.
Validation happens at the individual level. What matters is whether the model resonates as true for the person themselves.
These principles make clear why one-size-fits-all solutions rarely succeed. As van Zyl explains, the same experience can mean opposite things to different people. For one person, waking up at 3am may be a sign of insomnia. For another, it is a cherished time to join a Bible study group in another timezone and feel connected to their community.
Can AI Help?
If bottom-up approaches are the most accurate but also the hardest to scale, how can they be applied in practice?
This is where van Zyl’s work with artificial intelligence comes in. By analyzing thousands of personal narratives, his team has trained models to identify themes, cluster experiences, and even predict outcomes like burnout with surprising accuracy.
For example, language analysis showed that certain “high risk” words correlated strongly with burnout. Sentiment analysis revealed that the balance between personal demands and resources could explain additional variance. Taken together, these models could estimate burnout risk without relying on traditional surveys.
The goal is to create hyper-personalized assessments that capture what uniquely drives or detracts from each individual’s well-being.
Eventually, these assessments could generate equally personalized interventions, though van Zyl acknowledges that designing content to match each person’s profile remains a significant challenge.
Risks, Ethics, and Boundaries
As powerful as these tools can be, van Zyl and Soren emphasize the risks.
AI can deepen inequalities, erode uniquely human skills, and displace the personal connection that is central to care. There are also profound ethical concerns around data ownership, transparency, and consent.
Van Zyl stresses that individuals must remain at the center. People should know what is being tracked, why it is being used, and how long it will be stored. Without informed consent and clear ownership, even the most sophisticated models risk becoming exploitative.
Soren connects this to lessons from Sara Wolkenfeld’s episode, where she draws from Jewish teachings about the difference between rote work and sacred work. In the same way, we must decide what tasks we are willing to outsource to machines, and what forms of “sacred work” (such as, creativity, ethical discernment, human connection) we must hold onto ourselves.
Designing the Future of Work
Van Zyl believes AI should be used to augment, not replace.
It can handle pattern recognition, data analysis, and repetitive tasks, freeing people to focus on the parts of work that truly require human judgment and care. The challenge is not just technical but ethical: ensuring that these systems are designed intentionally, with safeguards to prevent harm.
As Soren reflects, the future could go two ways.
Left unchecked, AI might reduce work to hollow oversight of algorithms. But with thoughtful design, it could expand access to care, provide individualized support, and elevate the human aspects of work that matter most. He describes it as letting AI serve as an assistive tool, like a bionic arm that extends our capabilities, rather than a replacement for the uniquely human skills that give work meaning.
Key Takeaways
Top-down models of well-being overlook cultural and personal differences.
A bottom-up approach treats each person as unique, embedded in systems, and evolving over time.
AI and machine learning can help scale these insights, but they raise risks around ethics, dependency, and dehumanization.
Context and meaning shape how the same experience impacts well-being.
The future of work depends on using technology to enhance, not replace, what makes us human.
This conversation challenges assumptions about how we measure well-being and invites us to think critically about the role of AI in the workplace. By blending rigorous critique with a vision for what is possible, van Zyl offers both caution and inspiration for designing the future of meaningful work.