Sarah Martinez stared at the org chart on her laptop screen, the cursor blinking where five names used to be. In three weeks, her customer service department would transform from ten human agents to five humans working alongside five AI assistants. The efficiency projections looked promising on paper, but sitting in her office at 6 PM, she felt the weight of uncharted territory.

“The AI agents will handle routine inquiries,” the VP had explained during the announcement. “Your team will focus on complex cases and relationship building.” It sounded clean and simple in the boardroom, but Sarah knew management never was.

The First Challenge: Redefining Roles

Sarah’s first instinct had been to focus solely on managing her human employees—after all, you can’t have a one-on-one with an algorithm. But during the pilot week, she quickly realized this approach was failing spectacularly.

Marcus, one of her senior agents, stormed into her office on Tuesday morning. “Sarah, AI-Agent-003 is giving customers incorrect information about our return policy. I’ve had to fix three cases already today.”

“Did you report it to IT?” Sarah asked, falling back on her usual response to technology issues.

“IT says it’s working as designed. They told me to talk to you about ‘workflow optimization.’” Marcus made air quotes, frustration evident.

Sarah paused. This wasn’t a traditional HR issue or a technical glitch—it was something in between. She was managing a team where half the members couldn’t be coached in the conventional sense, but their performance directly affected her human employees’ success and job satisfaction.

The Revelation

That evening, Sarah had an epiphany while reviewing performance dashboards. She wasn’t managing 5 humans plus 5 separate AI tools. She was managing 5 human-AI partnerships—collaborative units where success depended on both components working in harmony.

She started thinking of each human employee as the “team lead” of a human-AI pair, but quickly realized this created its own problems. Jessica, her newest hire, felt overwhelmed by the responsibility of “managing” her AI counterpart when she was still learning the job herself.

A New Management Framework

Over the following weeks, Sarah developed what she privately called “hybrid management”—a framework that treated human-AI pairs as integrated units while recognizing their distinct needs.

For the humans, she maintained traditional management practices: career development discussions, performance feedback, emotional support during difficult customer interactions. But she added new dimensions: training on effective AI collaboration, helping them understand their AI partner’s capabilities and limitations, and mediating when human-AI workflows broke down.

For the AI components, she couldn’t provide motivation or career guidance, but she could ensure they had clear parameters, accurate training data, and feedback loops. She learned to read AI performance metrics the way she once read human behavioral cues, spotting when an AI agent was making systematic errors or when its responses were becoming too rigid.

For the partnerships, she focused on workflow design, ensuring handoffs between human and AI were smooth, that escalation paths were clear, and that the AI enhanced rather than hindered human decision-making.

The Weekly Dance

Sarah’s weekly team meetings became exercises in orchestrated complexity. She’d start with traditional human concerns—workload balance, professional development, team morale. Then she’d shift to AI performance reviews—accuracy rates, response times, customer satisfaction scores for AI-handled interactions.

But the most critical part became the “partnership reviews”—examining how well each human-AI pair was collaborating. When David reported that his AI partner excelled at data lookup but struggled with nuanced policy interpretations, Sarah worked with IT to adjust the AI’s escalation triggers while coaching David on better prompt engineering.

The Unexpected Discovery

Three months into the transition, Sarah made a surprising discovery during her quarterly review with the VP. The highest-performing pairs weren’t necessarily those with the most experienced humans or the most advanced AI configurations. They were the partnerships where the human had learned to think of the AI as a specialist colleague rather than a tool—setting clear expectations, providing context, and building on the AI’s outputs rather than simply accepting or rejecting them.

Emma, a mid-level agent, had started briefing her AI partner each morning about unusual customer situations or policy updates, treating it like she would a junior colleague. The AI’s contextual responses improved dramatically, and Emma’s job satisfaction increased as she took on more of a mentoring role.

The Ongoing Evolution

Sarah realized that managing a hybrid workforce wasn’t about applying traditional management techniques to humans while treating AI as sophisticated equipment. It required developing new skills: understanding AI capabilities and limitations well enough to optimize partnerships, designing workflows that leveraged both human creativity and AI efficiency, and helping human employees evolve their roles rather than simply being replaced.

She also learned to manage the emotional complexity of the transition. While productivity metrics were strong, some team members struggled with identity shifts—from being customer service representatives to becoming customer relationship specialists supported by AI. Others thrived in their new roles as human-AI team coordinators.

The New Reality

Six months later, Sarah’s department had found its rhythm. Her role had expanded beyond traditional people management to include something she called “ecosystem management”—ensuring that humans, AI, processes, and technology all worked together effectively.

She still had one-on-ones with her human employees, but these conversations now included discussions about AI collaboration strategies. She still reviewed performance metrics, but these now included partnership effectiveness scores alongside individual productivity measures.

Most importantly, she had learned that managing a hybrid workforce meant accepting that the traditional boundaries of management—human performance, technology deployment, process optimization—had blurred into something more complex but ultimately more interesting.

As she prepared for next quarter’s expansion to include more AI capabilities, Sarah reflected that the future of management might not be about managing humans or managing technology, but about managing the spaces between them where the real work increasingly happened.

Sarah Martinez closed her laptop and smiled. Tomorrow, she’d be interviewing candidates for a new position: Human-AI Partnership Coordinator. The job description had taken her three drafts to write, but she was finally learning to manage in a world where the org chart included entities that never got tired, never asked for raises, but still somehow needed leadership.

Authors commment

The story is set in a company transitioning its customer service team into human–AI pairings where each employee collaborates directly with an AI counterpart.

That scenario fits most closely with State 2: Collaborative AI:

• The AI agents are embedded in workflows but not autonomous decision-makers.
• Humans still hold responsibility, authority, and context, while AI handles repetitive tasks, lookup, and first-line responses.
• Management evolves into “hybrid management,” balancing human development and AI performance — classic State 2 dynamics.
• The discovery that the best performers are those who treat AI as a colleague rather than a tool underscores the partnership framing of this state.

It isn’t yet State 3 (Enterprise AI) because:

• The organization hasn’t been fully reconfigured around AI (no fluid teams or daily reassignments).
• AI isn’t in formal leadership roles (as in State 4).

Thanks

Special thanks to the 3x3 Institute for developing the AI State Model and pioneering the technologies and tools that enable progress toward the higher stages of AI-driven human achievement.