High-agency as the new operating system for AI-era work
BCG’s North America People & Organization lead Julia Dhar is putting a precise label on what many executives sense but struggle to articulate: as AI and automation redraw job boundaries, the differentiator is increasingly human agency—the capacity to self-direct, clarify ambiguity, and drive outcomes without waiting for instruction. BCG’s 2026 AI at Work Research frames the scale of the shift with two telling signals: 72% of employees report changed skill expectations, and 88% anticipate major upskilling needs over the next five years.
What makes the “high-agency mindset” especially consequential is that it is not merely a personality trait; it is becoming a workforce capability that determines whether AI adoption translates into measurable productivity and competitive advantage. In practical terms, high-agency employees:
- Clarify roles and decision rights when responsibilities blur
- Anticipate problems earlier and reduce operational surprises
- Prototype solutions independently, then socialize them with stakeholders
- Continuously improve workflows, including AI-enabled processes and controls
As organizations move from experimenting with generative AI to embedding it into core operations, the labor market signal is clear: initiative, judgment, and self-management are rising in value at the same time that routine execution is being commoditized.
The productivity paradox: AI saves time, but strategy determines value
One of the most revealing data points in the research is that 42% of workers say AI has freed up an average of one workday per week—a remarkable release of capacity. Yet two-thirds lack guidance on how to redeploy that time. This gap is where many AI programs quietly lose momentum: technology increases throughput, but without a clear reallocation plan, the “saved” hours can dissipate into low-value activity, fragmented experimentation, or simply a faster pace of the same work.
This is the modern productivity paradox in a new form. The constraint is no longer access to tools; it is organizational clarity and behavioral follow-through. High-agency employees bridge the gap by treating time as an investable asset—redirecting it toward higher-leverage outcomes such as:
- Customer-facing improvements (faster response cycles, better personalization, proactive retention)
- Quality and risk controls (prompt governance, evaluation harnesses, audit trails)
- Process redesign (eliminating handoffs, standardizing inputs, reducing rework)
- Skill compounding (targeted upskilling, domain depth, AI fluency tied to business goals)
A key technological implication follows: as AI absorbs routine tasks, individual accountability rises. Frontline contributors increasingly become the “operators” of AI workflows—configuring tools, validating outputs, and embedding guardrails. In effect, many employees become micro-managers of automated systems, responsible not just for what gets done, but for how reliably and safely it gets done.
Dhar’s emphasis on behavioral science—shaped by her work co-founding a behavioral lab—adds an important corrective to AI transformation narratives. Technical training alone rarely delivers consistent ROI. Organizations also need behavioral levers that make initiative repeatable: peer accountability loops, visible norms for experimentation, and performance systems that recognize unassigned but high-impact work.
Talent stratification, entry-level reinvention, and macroeconomic stakes
The workforce implications are not subtle. If 88% of employees foresee major upskilling needs, the market is primed for skill polarization: those who can continuously learn and self-direct will move into higher-value roles, while others risk being sidelined by automation-enabled redesign. Over time, that can intensify:
- Wage dispersion between high-agency, AI-fluent talent and routine-task roles
- A sharper war for talent focused on adaptable problem solvers
- Increased pressure on companies to prove that AI investments translate into productivity, not just activity
Perhaps the most underappreciated shift is occurring at the entry level. The traditional model—junior employees executing defined tasks under close supervision—is weakening. As AI tools handle first drafts, basic analysis, and repetitive workflows, entry-level hires are more likely to be asked to orchestrate AI outputs, measure outcomes, and manage exceptions. That is a fundamentally different talent pipeline: onboarding must teach not only tools, but also judgment, escalation discipline, and governance.
There is also a macroeconomic dimension. Economies that cultivate agency at scale may be better positioned to absorb AI-driven disruption without triggering prolonged skill shortages. A workforce trained to self-correct, redeploy time, and improve processes can reduce operational drag—supporting productivity growth that helps offset inflationary pressures tied to constrained talent supply.
What leaders must redesign: incentives, guardrails, and “agency infrastructure”
The research points toward a management challenge as much as a technology challenge: guidance is scarce precisely when autonomy is expanding. If two-thirds of employees lack direction on how to use AI-freed time, leadership’s role becomes setting strategic priorities and guardrails, not prescribing every step. The goal is to create an environment where initiative is safe, legible, and rewarded.
Several strategic moves stand out as “agency infrastructure” for AI-era organizations:
- Embed agency metrics in people analytics: track self-initiated projects, cross-functional problem solving, and measurable time-redeployment outcomes—not just task completion.
- Build internal agency accelerators: short-cycle programs where employees propose and pilot solutions to friction points, with seed resources and executive visibility.
- Publish AI time-redeployment playbooks: a curated menu of high-leverage activities aligned to business priorities, helping employees convert capacity into outcomes.
- Reskill with behavioral interventions: pair AI training with coaching on goal-setting rituals, feedback loops, and peer accountability—turning mindset into habit.
- Reward unassigned value creation: update KPIs and promotion criteria so process improvements, internal tools, and customer-driven enhancements are recognized, not treated as extracurricular.
The throughline is that AI is flattening hierarchies by necessity: when more employees are effectively designing and governing workflows, organizations must evolve toward clear decision rights, rotating leadership, and project-based pods that match the speed of change.
The companies that pull ahead will not be those with the most AI licenses deployed, but those that reliably convert AI-enabled capacity into better decisions, better customer outcomes, and faster learning cycles—powered by a workforce trained and trusted to act with agency when the playbook is still being written.
By
By
By
By
By
By
