Building a High-Agency Workforce
“High agency” here means people: curiosity, discipline, and initiative in real jobs under AI-intensified work, not AI agents.
For a while now I have felt conflicted when writing (or actually, nowadays it is “writing”) code. One moment I can conquer the world; the next, I need to pick up the pace just to keep up. My teammate called it a constant rush, and he was spot on. The irony is that AI was supposed to free up time, but instead that time gets filled immediately.
The February 2026 Harvard Business Review piece on how AI intensifies rather than reduces work describes the mechanics: with AI, the scope of work expands into areas that used to sit with others, involvement with work becomes more continuous throughout the day, and multitasking comes with heavy context switching. I recognize all three in my day-to-day. The article finally acknowledged the feeling, but the next question is what this means for organizations, and for the people they cannot afford to lose.
Drivers of AI work intensification
The article emphasizes enthusiasm: people reinvest gains because the work is engaging. Commentary on the HBR study captures this as follows: AI makes exploration addictive; because the feedback loop is instant; because the barrier between idea and execution has collapsed. It is argued that this is a manifestation of the Jevons paradox: Efficiency lowers the effective cost of doing things with the resource, but as a result it gets used more widely and more intensively.
That is one stream. The other is best captured under the term technostress. Technostress arises from individuals’ efforts to cope with evolving technologies and their changing cognitive and social demands. The five factors of technostress are as follows: ‘t-overload’ pushes users to work faster and longer; ‘t-invasion’ blurs the boundary between work and personal life; ‘t-complexity’ arises from technologies being too complicated; ‘t-insecurity’ stems from the fear of job loss to technology or more skilled individuals; and ‘t-uncertainty’ results from constant technological changes, necessitating ongoing education.
Developer burnout in the AI era
Enthusiasm and strain are not mutually exclusive; they run in parallel. But it matters for organizations to understand which one is driving behaviour. Stress-driven burnout comes with early warning signals such as complaints and disengagement. Passion-driven burnout is harder to detect because the person does not read as checked out. A subtler signal is overcommitment held together with ambivalence: high investment alongside conflicted or thinning desire. Organizations should monitor for both kinds of strain in their workforce.
Why high-agency professionals are a retention risk
The combination of technostress (largely caused by organizational factors) and AI enthusiasm (which is intrinsically rewarding) also creates a retention risk. For high-agency professionals, this imbalance quietly pulls toward independence. They experience the enthusiasm regardless of context, while the stress is environmental.
As argued in Shell Theory, it is agency more than experience that will separate people in the future. So high-agency is what organizations should recruit for. The same traits that make someone valuable (curiosity, discipline, initiative) also make them the most capable of building something on their own. And the stressors they face at work are precisely the ones that disappear when they leave. This makes high-agency professionals the most likely to leave and the hardest to replace, because independence keeps the enthusiasm and drops the stress.
Designing for retention
Intensification Guardrails
As a starting point, organizations should implement guardrails that align with the three practices from the HBR article. The same guardrails that reduce technostress also shape how enthusiasm is spent. That combination retains high-agency personnel:
| Practice | For the Individual | For the Organization |
|---|---|---|
| Take intentional pauses | Protect gaps that used to exist before micro-gaps filled with iteration. | Normalize uninterrupted focus blocks. |
| Sequence your activities | Cut context-switch tax by ordering work deliberately. | Set clear priorities and limit concurrent and competing initiatives. |
| Connect with co-workers | Intensity in isolation skews judgment; social context is a resource. | Encourage co-presence and collaborative rituals. |
Invest in high-agency juniors
As argued in Shell Theory, everyone below the flatline produces the same output regardless of skill. The temptation is to stop investing in juniors entirely, because they cannot surpass what autonomous agents can (or will) give us. This is a mistake.
The right investment is targeted: juniors who show agency (curiosity, discipline, initiative) and strong cultural fit. Agency accelerates their path into the amplification zone, and cultural fit makes them more likely to stay once they get there. The result is a durable core of capable people who are also committed.
Autonomy as retention
If there is a pull toward independence, the counter-move is to make the organizational environment feel more independent. That means high autonomy, low bureaucracy, and clear ownership. Not unlimited freedom, but structured autonomy: freedom within explicit boundaries.
Conclusion
Building a high-agency workforce is not just about hiring for agency. It is about designing an environment that retains the agency it recruits: guardrails against intensification, targeted investment in high-agency juniors, and structured autonomy that makes staying feel like independence. The rush my teammate described is not going away. AI collapses the space between idea and execution, and the people who thrive in that space will keep filling it. Whether that energy is spent within the organization depends on the environment around it.
For individual contributors wondering where they stand, the Shell Theory self-test can help surface which zone you’re operating in.