Why the skills that make us irreplaceably human are now your organisation's most valuable asset — and why most organisations are failing to develop them at exactly the wrong moment.
The conversation about AI and work has been dominated by a single question: which jobs will it take? This is the wrong question. The more useful question — and the one that organisations need to be answering now — is: when AI handles the analytical and procedural work, what does that leave? What kind of work becomes more valuable?
MIT Sloan researchers Isabella Loaiza and Roberto Rigobon published their EPOCH framework in 2024, with findings released publicly in March 2025. Analysing task data across all US occupations between 2016 and 2024, they found a measurable, sustained shift toward more human-intensive work. New tasks introduced to the labour market in 2024 carried significantly higher scores for human-intensive capabilities than the tasks that existed before. Jobs intensive in human capabilities showed stronger employment growth, higher hiring rates in 2024, and more favourable projections through 2034.
This is not a prediction about the future. It is a description of a shift that has already occurred.
The 73% figure from Ransbotham et al.'s MIT Sloan / BCG research is worth dwelling on. Nearly three quarters of digital transformation projects fail to deliver their expected value — not because of technical failure, but because of human resistance, low trust, and poor adoption. The technology works. The people don't follow.
The organisations that make AI transformation work are the ones that invest in the human infrastructure alongside the technical infrastructure. Leaders and managers with the empathic capacity to take people through change — to read the room, sense resistance before it becomes obstruction, create conditions in which people feel safe enough to adopt something unfamiliar. This is not a soft capability. It is the capability that determines whether the hard investment pays off.
MIT Sloan's EPOCH framework names five groups of capabilities that remain stubbornly human in the face of AI development. Together they describe the profile of the human worker whose value increases as AI becomes more capable. The researchers are explicit: "We deliberately don't call these soft skills. It is much harder to teach a person these critical human skills and capabilities."
| Capability | What It Means in Practice | Why AI Cannot Replicate It |
|---|---|---|
| Empathy ↑ | The capacity to understand, feel, and respond to the inner states of other people. In professional contexts: the quality of attention that makes people feel genuinely heard. | Requires physical co-presence and neural mirroring between embodied nervous systems. AI can detect emotion patterns but cannot share them. |
| Presence | Physical and relational presence — the capacity to be genuinely available to another person. In management: what makes a one-to-one feel like a conversation rather than a transaction. | Requires a body in a room. The quality of presence is conveyed through micro-signals — posture, breath, attention — that are both physical and involuntary. |
| Opinion | Judgment, ethics, and the capacity to navigate open-ended situations where there is no correct answer, only a defensible one. In leadership: taking a position and explaining it. | AI can synthesise existing positions but cannot hold genuine accountability or exercise ethical judgment in novel situations with real consequences. |
| Creativity | The imagination to generate genuinely new ideas — connections that have not been made before, solutions that draw on lived experience and aesthetic judgment. | AI can recombine existing patterns at scale. It cannot experience the world — the source from which genuine creative insight emerges. |
| Hope | The capacity to sustain motivation and meaning in conditions of uncertainty. In organisational terms: what keeps teams functional under pressure when outcomes are unclear. | Hope is relational and embodied. It is communicated through physical presence, tone, bearing. It cannot be generated by a system that does not experience uncertainty. |
The WEF's finding that empathy shows zero substitution potential is specifically a finding about the physical dimension: the human body in a room, attending to another human body, in the way that only co-present nervous systems can. The pattern across the three empathy dimensions is striking — and points directly at where development investment needs to go.
"Empathy is not a warm fuzzy feeling. It is a physical skill — raw, fast, instinctive. It is what happens when your body tunes into someone else's. And like any instrument, you can learn to play it."Stuart Nolan · How to Train an Empath, 2025
The evidence that empathy is becoming more economically valuable is robust and growing. The evidence that organisations are responding to this by investing more in empathy development is not. The pattern, in most organisations, runs in the opposite direction.
The CIPD estimates UK organisations spend over £5.7 billion annually on management training. The majority is directed at cognitive and technical capabilities. The fraction directed at developing the empathic capabilities that determine management effectiveness in human relationships is small — and declining. When AI handles the cognitive tasks, what remains of management is almost entirely relational. The manager's job becomes the work that requires empathy, presence, and genuine support through change.
Even where empathy development does take place, it tends to operate at the level of awareness. Programmes built around personality profiling, emotional intelligence frameworks, and reflective practice are genuinely good at producing insight. What they do not reliably produce is a manager who responds differently at 9am on a Monday when a team member is in distress and a deadline is pressing. The management relationship is played out in real time, under pressure, in the body — not in a reflective journal.
In each case, the context included significant AI or digital transformation pressure. The intervention was not additional technology, additional process, or additional cognitive training. It was investment in the quality of physical human attunement between people who needed to work together under conditions designed to make that harder.
"The moment someone gets it — when doubt flips to WTF! — is still the best part of my job."Stuart Nolan · How to Train an Empath, 2025
Download the complete white paper for the full EPOCH framework analysis, WEF substitution data, and practical guide to building the AI-era case for physical empathy investment. Or book a free 30-minute discovery call.