Joanne Chen of Foundation Capital presents findings from interviews with 25 companies about how AI is driving structural organizational change, not just individual productivity gains. The research reveals dramatic shifts: a 120-person engineering team planning to shrink to 25, expert-to-generalist ratios moving from 1:6 to a target of 1:100, and traditional roles like product, engineering, and design collapsing into fewer hybrid positions. The piece draws an analogy to factories switching from steam to electricity — the real gains came not from swapping power sources but from redesigning the entire system around the new technology.

The article identifies four human roles that remain essential in the AI-augmented org: chief accountability officers (executives who own outcomes), systems architects (who design human-agent collaboration), relationship experts (who maintain human-to-human trust), and validators (who review and approve agent outputs). Validator demand is expected to follow a bell curve, peaking in 2–4 years before declining as AI systems self-improve — but a critical "one-generation problem" emerges: if agents handle all junior work, the next generation of experts may never develop the foundational skills needed to validate AI output.

Chen maps four company archetypes along digital-vs-physical and product-vs-service axes, each facing distinct reorg patterns. Digital services firms face the most immediate pressure, as agents increasingly are the product. Seven startup opportunities emerge from these shifts, including tools for human system architects (like Arize), validator platforms (like Turing), agent-native infrastructure (like Sixtyfour), and AI-native services firms (like ConverzAI and Tessera Labs). The piece concludes that systems thinking — the ability to design workflows across humans and agents — is the most valuable and hardest-to-develop skill in the emerging organization.

Microsoft's 2026 Work Trend Index report, drawing on trillions of anonymized Microsoft 365 productivity signals and a survey of 20,000 AI-using workers across 10 countries, argues that AI and agents are fundamentally expanding human agency at work — but most organizations aren't built to capture the value. The report introduces the concept of a "Transformation Paradox": employees are ready to reinvent how they work with AI, but organizational metrics, incentives, and norms continue to reinforce legacy ways of working. Only 19% of AI users occupy the "Frontier" zone where both individual capability and organizational readiness are high, while the largest share sits in an "emergent" middle ground of misalignment.

The research identifies Frontier Professionals — 16% of AI users who deploy agents for multi-step workflows, routinely redesign processes, and build shared AI standards — as disproportionately valuable. These users prioritize critical thinking and human judgment, intentionally doing some work without AI to keep skills sharp. Crucially, the report finds that organizational factors like culture, manager support, and talent practices account for more than 2x the reported AI impact of individual effort alone, underscoring that the biggest lever for AI value isn't hiring better people but creating the right environment.

The report calls on organizations to become "Learning Systems" — firms that capture signals from agent-driven work and encode them into shared routines, building what Microsoft calls "Owned Intelligence" that compounds over time. It outlines coordinated roles for employees, leaders, IT, and security in building an evaluation infrastructure for agent operations at scale. Manager behavior proves especially impactful: when managers actively model AI use, employees report a 17-point lift in AI value, a 22-point lift in critical thinking, and a 30-point lift in trust in agentic AI.

IESE professors Enrique Ide and Eduard Talamas present a framework for understanding how AI may reshape the knowledge economy, focusing on two critical variables: AI autonomy and computing power availability. Their research distinguishes between non-autonomous AI, which always requires human collaboration and broadly enhances productivity, and autonomous AI, which can operate independently and creates more complex labor dynamics with clear winners and losers.

The framework further differentiates between basic autonomous AI (capable of routine knowledge work) and advanced autonomous AI (capable of specialized problem-solving), each producing distinct organizational effects. Basic AI displaces workers from routine roles into more knowledge-intensive positions, leading to smaller, more centralized firms. Advanced AI does the opposite — it reduces the knowledge content of human work and creates larger, more decentralized organizations. Crucially, the most knowledgeable workers always benefit from AI regardless of its sophistication, as they can leverage it to amplify their expertise.

The question of whether AI will destroy knowledge jobs ultimately hinges on computing power relative to AI's potential applications. If compute remains scarce relative to demand, no one's labor becomes redundant. But if computing power becomes sufficiently abundant, autonomous AI could render workers whose knowledge falls below AI's level unemployable — a scenario with profound implications for both individual career strategy and organizational design.

Harvard Business Review highlights how AI is accelerating product development and placing increased demands on marketing teams, yet most organizations remain hampered by siloed workflows and coordination-heavy structures designed for an earlier era. The post points to a growing trend of companies fundamentally rethinking their organizational models to keep pace with AI-driven speed and complexity, suggesting that legacy operational frameworks are becoming a bottleneck rather than a foundation for growth.

Microsoft's 2025 Work Trend Index introduces the concept of the Frontier Firm — a new organizational model built around on-demand AI intelligence and hybrid human-agent teams. Drawing on survey data from 31,000 workers across 31 countries, LinkedIn labor trends, and Microsoft 365 telemetry, the report argues that intelligence is becoming an abundant, affordable resource that fundamentally rewires how businesses operate. A critical capacity gap emerges from the data: 53% of leaders say productivity must increase, yet 80% of the global workforce reports lacking sufficient time or energy. Employees face 275 interruptions per day during work hours, and after-hours work continues to climb. AI agents are positioned as the solution, with 82% of leaders expressing confidence they'll use digital labor to expand workforce capacity within 12–18 months.

The report outlines three phases in the journey to a Frontier Firm: AI as assistant, agents as digital colleagues on teams, and agents running entire business processes with human oversight. A key concept is the Work Chart — replacing traditional org charts with dynamic, outcome-driven teams powered by agents. The report also introduces the idea of the agent boss: every employee managing AI agents as part of their role, from entry-level workers to executives. Leaders currently outpace employees in agent readiness across all seven indicators measured, creating a gap that organizations must close through training and cultural change.

Among the 844 Frontier Firm leaders identified in the survey, 71% say their company is thriving compared to just 39% globally, and they report significantly less fear of AI replacing their jobs. Nearly half of all leaders (45%) say expanding team capacity with digital labor is a top priority, while 78% are considering hiring for AI-specific roles. The report emphasizes that AI literacy is the most in-demand skill of 2025 according to LinkedIn, and urges organizations to move beyond pilots to broad-scale AI deployment, set human-agent ratios for different functions, and invest heavily in reskilling their workforce.

Eric Jing shares the third installment in his "Seeing AGI" series, focusing on the concept of building AI-native teams. He argues that creating such teams goes beyond simply adopting new tools — it requires a fundamental rethinking of how organizations structure, scale, and execute their work. Drawing on nearly 20 years of experience across Search, Hardware, and AI, Jing positions this as a critical shift in organizational design rather than a mere technology upgrade.

The post appears to be a longer-form piece linked from X (formerly Twitter), suggesting a deeper exploration of what it means to operate as an AI-native organization. This is part of an ongoing series where Jing examines the path toward AGI from practical, operational perspectives.