Higher education leadership for an agentic workforce

The five tiers of managing AI labor

By Jennie Wong, PhD

Colleges and universities are moving forward with building and deploying AI agents across their institutions, making some traditional frameworks for leadership more relevant than ever before.

As higher education institutions continue to build and mature their data and AI capabilities, the shift from generative AI to AI agents represents a phase change from “AI as tool” to “AI as digital labor.” And while IT decision makers have their hands full with evolving policies, platform decisions, and use case prioritization, there is also a need to connect this new AI workforce with some time-honored frameworks for leadership.

Up to this point, the new wave of AI has been understood as a revolutionary technology with incredible potential, including the personalization of learning experiences, faculty support, research data analysis, operational efficiency, and much more. Currently, AI can be tapped as a specific application on your phone, or an institutionally sanctioned environment with a choice of models, or as an embedded feature in our virtual meetings, collaboration tools, and just about every point solution and enterprise platform.

But even as user adoption is catching up to the promise of this wave, the next wave is already visible on the horizon. Beyond the current crop of pilots and proofs of concept, there is a multi-agent future that will be here sooner than you might think.

And this multi-agent future will shift our understanding of AI as a solution or feature to something more akin to our human coworkers. The multi-agent future will give each and every one of us a digital army to command, and that near-future scenario will require every employee to become a leader.

Luckily, when we examine our traditional frameworks for leadership, they give us meaningful ways to conceptualize, develop, and measure our effectiveness when leading an agentic workforce.

The five tiers of managing digital labor

We know that leading humans starts with self-management, followed by basic supervision. Supervisors graduate to middle management, where they provide the connective tissue between executives and the front line. A subset of these managers ascends to the executive level, where they are tasked with setting strategic priorities or even spin-off their own entrepreneurial endeavors.

Developing the skills to lead AI agents can follow this same blueprint.

Tier 1: Self-management

Self-management is the foundation for leading others, regardless if they are flesh and blood or bits and bytes.

This is an area where many people are finding initial success, leveraging AI to keep themselves organized, efficient, and on track. This can look like a conversation with AI to clarify your goals or brainstorm tactics, asking for help to triage an overflowing inbox, or running a draft through AI to spot the logical weaknesses.

This foundation is created through increasing engagement, from small actions to habits to personal workflows, and results in a unique body of experience with AI that informs one’s approach to AI leadership in the subsequent tiers.

Tier 2: Supervision

Supervision for AI agents involves assigning discrete tasks and then inspecting for accuracy, quality, and thoroughness. If AI outputs are not reviewed, there is a risk of confident-sounding hallucinations where the user is flattered, sources are fabricated, and conclusions are untrustworthy.

Just as with human labor, good AI supervisors actively build their risk-calibrated approach to task review. What are the areas and actions where your AI agent is well-grounded and well-guarded against error? Which of its outputs is most likely to contain fabrications? Where does it make sense for a human to validate that the task has been done correctly?

Strong supervision is understanding that work can be delegated to an AI agent but accountability cannot.

Tier 3: Middle management

Middle managers have traditionally been the unsung heroes who translate strategic priorities into clear guidance for the front line. This remains true as our teams are expanded to include AI agents.

For example, suppose the Office of University Advancement has set an ambitious target for the annual campaign. How can this priority be converted into meaningful direction for an agentic workforce? Good managers operating in this tier will look for opportunities to adjust the work of AI agents to create the desired outcome. Maybe the donor research agent can expand its target list to include family members of alumni, or perhaps the impact reporting agent can increase personalization with social media content. Additionally, multi-agent orchestration could be enhanced to create and leverage deeper insights across donor types.

Tier 4: Executive leadership

Executive leaders in higher education articulate the short- and long-term priorities that reflect the institutional mission. They lead large swaths of the institution and are responsible for providing vision and strategy for areas such as learning and teaching, research and innovation, student success, and community impact.

Thinking like an executive will become a necessary skill for a broader range of leaders in an agentic future, where the work of multiple agents needs to ultimately ladder up to the most important success metrics.

Is the AI workforce effectively deployed to improve access, affordability, persistence, career readiness, and lifelong engagement? Are AI agents supporting the work of scholars and researchers as they push the frontiers of human knowledge? And are humans in the community experiencing a positive impact from the sum of university operations?

Tier 5: Entrepreneurial vision

Beyond the C-suite, there is a tier of leadership that agentic labor puts within reach. Though not everyone will become a founder of their own college or university, the ability to create something new becomes possible for all.

This could mean a new degree program, a specialized capability within an existing function, an experimental model for corporate partnerships, or any innovative endeavor that supports the mission. Strategy, planning, research, and execution constraints will be loosened to an unprecedented degree, giving all human team members the ability to imagine and explore what could be.

With access to an AI workforce, there will be no limits when pursuing higher education’s purpose.

Three takeaways for higher education leaders

1. An agentic workforce makes leadership a universal skill. As our current teams rapidly evolve to include AI workers, everyone becomes a leader, and we have the opportunity to develop skills across all five tiers of managing digital labor.
2. The five tiers of agentic leadership work together. Self-management is the bedrock, and supervision is the foundation. Increasing fluency in leading AI agents unlocks important middle management capabilities, executive thinking, and opens up possibilities for entrepreneurial vision.
3. The future of AI is a leadership opportunity. The scope of AI transformation goes beyond technology and will impact a broad range of organizational processes, including how we define the human role in managing this powerful engine.

The agentic workforce is already a reality, and just like the human workforce, it requires a range of leadership competencies to realize its potential. Ideally, this progression of leadership will direct the unique strengths of AI agents to the benefit of our institutions and people.

And while a machine may be faster or more thorough than a human, it only renders what the heart and mind guide it to create. To some, that’s artistry. To others, that’s leadership.

*Slalom is a fiercely human business and technology consulting company that leads with outcomes and teams with leaders, bringing more together.*