From managing people to managing AI: The leadership skills everyone needs now | Julie Zhuo (Facebook VP, Sundial CEO, The Making of a Manager author)

When managers become builders: a new script for leadership in an AI age

Julie Zhuo has one foot in the old world of high-scale product design and another in the new world of AI-infused decisionmaking. Her arc—from long-time design lead at a global social platform to founder of an AI start-up focused on automated analysis—creates a rare vantage point: she sees both the muscle and the machine, the intuition and the instrumentation. The practical tension she describes is deceptively simple: technologies are changing how work gets done, but the social work of organizing people, clarifying purpose, and sustaining growth demands steady, human judgment.

Redefining roles: builders instead of job titles

Zhuo argues for dissolving the rigid lines between designers, engineers, product managers, and data scientists. Where teams once required narrow specialists to be hired for each piece of a puzzle, modern tools allow individuals to become more generalist and capable—what she calls "builders." The implication is not that specialization disappears, but that teams can be smaller, more nimble, and composed of people who take ownership of outcomes rather than lanes. This shift reshapes hiring, accountability, and how work is divided: fewer silos, more cross-functional craft.

Manage outcomes, not activity

One of Zhuo's recurring prescriptions is deceptively tactical: be crystal clear about the outcome. Whether directing people or assembling agentic systems, the manager's job is to define success. She urges leaders to boil fuzzy ambitions down into objective criteria so that teams—and AI models—can be evaluated against the right target. That clarity reduces wasted motion, converts vague hopes into testable hypotheses, and enables a lean feedback loop that scales faster than process.

Diagnose with data, treat with design

This aphorism captures a critical synthesis. Data, Zhuo insists, should illuminate what’s happening and surface where the opportunity or problem lies. Design remains the creative, generative force for solutioning. The two together form a disciplined cycle: use analytics and instrumentation to discover root causes, then apply creative craft to prototype and iterate. That distinction helps teams resist the mistaken belief that analytics will hand them perfect answers—numbers diagnose; humans design.

Instrumentation, observability, and the paradox of fast growth

Rapid expansion creates a paradox: companies can amass users and revenue before they’ve built the logging, KPIs, and measurement plumbing that reveal why growth stalls. Zhuo emphasizes that many modern high-growth teams are operating on instinct until the first plateau. When growth slows, the absence of observability becomes painfully obvious. Her view reframes analytics not as a luxury but as a resiliency mechanism: instrument early, invent new methodologies for conversational and intent-driven data, and treat metrics as reflective tools rather than final verdicts.

Orchestrating models: assemble the right toolset

Models are personalities with distinct strengths, and leaders will need to learn the art of building an "Avengers" of AI systems—choosing agents and tools for specific tasks. The manager’s role becomes partly orchestration: decide which model should do which job, provide the context that enables good answers, and write the evaluation criteria that measure effectiveness. As Zhuo notes, better models may progressively handle more of the operational work, but humans still supply goals, judgment, and long-term priorities.

Feedback, self-knowledge, and the willow tree metaphor

Timeless management practice—managing yourself before managing others—remains central. Zhuo revisits familiar managerial advice with contemporary force: develop self-awareness of strengths and weaknesses, cultivate a feedback habit that accelerates weekly improvement, and hold a stable conviction while remaining remarkably flexible. Her willow tree metaphor—sturdy roots with flexible branches—captures a leadership posture that emphasizes steadiness amid rapid change.

Practical culture shifts: feedback as daily practice

To get better faster, teams need frequent, candid calibration. Zhuo recommends opt-in feedback cultures where colleagues explicitly agree to give and receive reflection often. That practice shortens learning cycles and prevents large misalignments from calcifying into personnel problems. Feedback, when framed as generosity and paired with sincere intention, becomes a multiplier for individual growth and team resilience.

Personal life, playful uses of AI, and emotional training

Outside the workplace, Zhuo demonstrates how AI can amplify human connection and parental imagination—creating personalized parody songs for a child or a talking raccoon companion—while also cautioning against using technology as an emotional shortcut. The deeper lesson she offers parents (and leaders) is to teach emotional regulation: the ability to sit with discomfort, notice bias, and choose hard growth over passive comfort remains essential when tools can now smooth so many frictions.

Concluding thought: judgment as the enduring comparative advantage

The era of agentic systems and flattened org charts shifts who does what, but it doesn’t erase the work of meaning-making. Clear metrics, better tooling, and more capable models change how decisions get made; they do not remove the need for human judgment. Where machines can accelerate skill acquisition and scale execution, managers who learn to set aims, curate instruments, and cultivate emotional stamina will convert technological capability into enduring value. The future will reward builders who know which problems to solve and why—they will be the stewards of both possibility and purpose.