Managers are unprepared to lead a human-agent hybrid workforce

Enterprise AI adoption has already reached saturation: 88% of organizations now use AI regularly in at least one business function, up from 55% in 2023. Most of this growth reflects the increasing number of employees emerging as large language model power users. But a more consequential shift is also taking place — one that poses a real challenge to the traditional role of a manager.

The deeper change is the emergence of true human–agent collaboration. Sixty-two percent of organizations are now experimenting with AI agents, with 23% already scaling agentic systems in at least one function. When an agent drafts a proposal, negotiates terms or triages customer requests without waiting for human approval, the nature of work itself changes. The employee’s role shifts from executor to orchestrator, not all that unlike a management role — just with AI “reports” instead of human ones.

In some cases, employees may explicitly hand off tasks to autonomous systems. In others, the shift will be gradual — AI suggestions give way to AI actions; human oversight shifts mainly to exception-handling. Leaders need new frameworks for this spectrum, which means rethinking foundational assumptions about expertise, collaboration and career development.

Domain expertise becomes a negotiable commodity

Leaders will need to decide whether to prioritize agent facility over domain expertise. Companies risk creating workforces skilled at directing agents but rusty in their own abilities.

This tension is most notably visible already in legal services, where AI adoption surged from 19% to 79% between 2023 and 2024 — the fastest shift of any profession. Over 95% of legal professionals now expect generative AI to become central to their workflow within five years. When AI can handle routine contract review with 94% accuracy, what skills should junior associates still master?

Leaders who reward skillful agent orchestration without insisting workers retain foundational knowledge may gain short-term efficiency but create brittle teams that can’t function when systems fail or edge cases emerge. The leadership challenge is distinguishing what knowledge remains essential from the competencies that can be safely delegated.

Collaborative culture demands intentional design

As agent-mediated work becomes the default, team collaboration won’t disappear, but it may quietly erode. When each team member can get instant, tailored feedback from AI, the friction of scheduling meetings and waiting for colleagues starts to feel unnecessary. Why sit through a brainstorming session when an agent can help iterate right now? The substitution happens gradually, one skipped conversation at a time.

This risk is acute in creative fields, where 60% of marketers now use AI daily, up from 37% a year ago. Shared creative culture relies on that essential glue of connected work. To ensure AI accelerates rather than erodes it, managers will need to become intentional architects of when and how their teams think together.

Career development pivots toward strategic thinking

As agents absorb execution-layer work, traditional advancement metrics — technical precision, production speed and operational excellence — may carry less weight than they once did. Leaders will need to re-weight the value of vision, judgment and orchestration over efficient execution (which AI can increasingly handle on its own).

This shift is already underway. Research on customer support teams using AI found productivity increased 14% on average, but gains were most pronounced — up to 34% — for less experienced workers. The execution gap between junior and senior staff is narrowing, at least in terms of sheer output. What differentiates them now is judgment, strategy and decision-making under ambiguity.

Managers will face their own identity crisis

Managers who built careers on technical expertise may feel exposed as AI makes those skills more accessible to everyone — any employee with a good prompt can now punch above their weight. But the same technology can also absorb the administrative drag that eats up a manager’s day: meeting prep, status tracking and the like. What remains is the work that actually differentiates a good manager: building trust, navigating ambiguity, having difficult conversations and sensing when someone’s struggling before they say so. Somewhat paradoxically, the same technology that creates a new learning curve for managers can also enhance their own effectiveness.

Leading through the agent transformation

To prepare managers for this new era of work, talent leaders will need to decide where they stand on these key issues: What performance metrics are acceptable when agents share the work? What cultural shifts are we willing to accept in exchange for efficiency? The answers will vary by company culture and objectives. Still, every organization will have to tackle this sooner rather than later, or let the unruly momentum of adoption dictate their direction.

Policies will need updating as the technology evolves. While still squarely in the agentic hype cycle, there will likely be bumps on the road to mature adoption. Gartner is predicting that upwards of 40% of agentic AI projects will be canceled by 2027 due to unclear value or inadequate controls. It’s a moment that calls for clear leadership and nimble adaptation.

The companies that navigate this well will be the ones that get specific about what good management actually looks like when teams go AI-hybrid and build their talent strategy around those commitments.

Partner insights from BetterUp