Half your team develops through coaching. The other half earns trust through verification. The NEXUS Leadership Model™ is a research-grounded framework for leading both — built for managers and individual contributors.
Situational Leadership II. GROW. Tuckman. DiSC. All excellent — and all built on assumptions that fail the moment an AI agent joins your team: that team members have motivations, develop through experience, carry legal accountability, and respond to rapport. AI agents have none of these. Yet they're doing real work on your team, right now.
Manage an AI agent like a motivated human, and you'll over-trust it — catastrophically. Manage a human like a system to be specified, and you'll destroy the psychological safety that makes people great. Leaders need one framework that does justice to both.
NEXUS gives every contributor — human or AI — a stage, and gives the leader a matched behavior set for each. Humans progress through development: competence and commitment grown through coaching. AI agents progress through verification: autonomy earned through empirically demonstrated reliability. Click any stage to see exactly how to lead it.
Most hybrid teams don’t stumble inside either track. They stumble where the work changes hands. Here is one decision, worked through.
Sam spot-checks 20 extractions against the source. Two are misattributed. The agent stays at A1 for this client: full review continues, the extraction instructions get tightened, and the check runs again before review drops to sampling.
Sam’s catch gets named in front of the team. At H3, visible recognition is part of the job.
Leading as an individual contributor? Put yourself in Sam’s seat. The analyst directing the agent makes the same three moves — that’s the IC Lens.
In a pre-registered NBER experiment with 249 leaders (Weidmann, Xu & Deming, 2025), leadership skill with AI agents correlated ρ = 0.81 with causal leadership impact on human teams — and still 0.69 after accounting for the leaders’ own task skills. The leaders who did well with both did the same observable things: asked more questions, took more conversational turns, and used more "we" language. NEXUS builds the first two behaviors on those findings and adds three drawn from leadership and AI-governance practice — together, the five universal behaviors of hybrid leadership.
Ask more questions; issue fewer directives.
Lead with "we" — collective language outperforms directives.
State the what and the why — always.
Evaluate results, not process.
Match oversight to demonstrated reliability — neither more nor less.
If you're already an effective leader of people, you have more of what it takes to lead AI agents than you think. The correlation runs both ways — and it's the single most reassuring finding in hybrid-leadership research.
Stages describe individuals. Modes describe moments. NEXUS leaders shift among four team-level modes based on the challenge in front of them.
Design roles, boundaries, and success criteria — written expectations for humans, constraint architecture for agents. Ask: does every member, human and AI, have a documented understanding of role, authority, and limits?
Optimize the handoffs. Facilitate collaboration and recognition for humans; tune workflows and oversight calibration for agents. Ask: is every contributor at their optimal level, with seamless handoffs between them?
Provide clarity and safety in ambiguity. Sensemaking and change leadership for humans; human-in-the-loop escalation and risk containment for agents. Ask: can this team act effectively in conditions I can't fully specify?
Delegate broadly, govern vigilantly. Career investment and innovation latitude for humans; drift detection and audit discipline for agents. Ask: am I giving real latitude to those who earned it — while keeping the vigilance trust requires?
If you direct an AI agent — assign it work, evaluate its output, decide what it may and may not do — you are exercising leadership, whatever your title says. Skills once reserved for managers are now baseline professional competencies. Most organizations haven't noticed. The ones that have will out-execute the ones that haven't.
That's why NEXUS is one framework with two seats. The Agent Track is yours as-is: every agent you direct, you lead. And the Human Track flips to your perspective — switch the explorer above to the IC Lens to see each stage from your seat: know where you are, ask for what you need, and make your growth visible.
| Responsibility | The Old Reality | The New Reality |
|---|---|---|
| Goal Setting | "Complete this task as instructed." | Define outcomes clearly enough for an AI to act on them accurately. |
| Delegation | "That's my manager's job." | You decide what to delegate to AI, what to retain, and where the handoff lives. |
| Evaluation | "My output is my output." | You evaluate AI-generated work before it becomes your output. |
| Escalation | "Not applicable to my role." | You recognize when AI output requires human review — and escalate. |
| Accountability | "I own what I personally produce." | You own every output you sign off on, regardless of what produced it. |
| Communication | "I talk to teammates in meetings." | You design instructions precise enough to direct autonomous action. |
The L&D implication: prompt design, output evaluation, and escalation judgment are not IT training topics. They are leadership competencies — and NEXUS treats them that way at every organizational level, in both seats.
Eight questions. Two minutes. A profile of your human-track and agent-track leadership instincts — and the one behavior to work on first. Fair warning: the obvious answer isn't always the right one.
NEXUS synthesizes peer-reviewed and institutional research on hybrid human–AI teaming. Core sources:
Workshops for people leaders. IC programs for everyone who directs an AI agent. Executive briefings for the leaders deciding what comes next. Delivered through your existing L&D infrastructure.