.png)
AI can analyze information, make recommendations, and execute complex workflows in minutes. Yet work that takes minutes to execute can still take days to move through an organization, waiting for reviews, decisions, and approvals.
Working with many organizations on AI, I keep seeing the same pattern. AI is transforming what we can accomplish, but our organizations are still designed around how we used to accomplish it. Increasingly, the bottleneck isn't the ability to execute work. It's the organization surrounding it.
We Built Organizations Around Constraints AI Is Now Challenging
For decades, we designed organizations around the limitations of how work got done. Information was fragmented and often lived in people's heads, so we created management layers to collect, interpret, and communicate it. Expertise was concentrated in specialized departments, so work moved through functional handoffs. Execution was expensive and time-consuming, requiring detailed planning, approvals, and controls. And because capacity depended heavily on people, growth meant adding headcount, management layers, and coordination.
These structures were logical responses to an era when efficiency, predictability, and control were essential to scaling a business.
But AI is challenging the assumptions behind that design. Information can now be analyzed continuously across functions. AI agents can coordinate and execute tasks that previously required multiple specialists. Capacity can expand without a proportional increase in headcount, while the ability to adapt quickly becomes increasingly valuable.
Consider how businesses operate: weekly status meetings, monthly reviews, quarterly planning, and annual budgets. These routines were designed when information had to be gathered, summarized, and reviewed periodically. AI agents can work continuously, but their outputs often enter the same approval queues built for an earlier era.
AI operates continuously, but organizations still rely on decision-making structures designed for a slower era.
Redesigning decision architecture is foundational to building an AI-enabled organization. This means fundamentally rethinking who or what has the authority to make decisions, at what level of autonomy, under which conditions, and with what accountability.
Rather than routing every decision through traditional management hierarchies, routine, low-risk activities can proceed autonomously within defined guardrails. Consequential decisions retain appropriate human authorization, while clear escalation and override mechanisms address exceptions, uncertainty, and elevated risks. The objective isn't to eliminate human oversight, but to design decision-making and governance for an environment where people and AI work together.
Without this shift, even the most capable AI systems remain constrained by the very decision structures they were meant to improve.
Stop Automating Silos. Break Them by Redesigning Work Around Outcomes.
Most organizations ask how AI can make their existing processes faster. The more important question is: If we designed the business from scratch today, knowing what AI can do, would we organize work the same way?
Today, we still treat functional departments as the primary building blocks of how work gets done. People are organized by expertise, work is divided among departments, and management hierarchies coordinate activities and resolve dependencies. Each function owns its part of the process, but accountability for the complete business outcome often gets fragmented along the way.
Consider a customer onboarding process involving sales, operations, finance, and technology. Each department completes its responsibilities and passes the work to the next. AI may make every department more productive, but if the same handoffs, approval chains, and organizational boundaries remain, the customer may see little improvement.
We explored this challenge in our white paper, The Transformation Gap, in Process Architecture: Automating the Past.
The real transformation isn't making departments more productive. It's making business outcomes, rather than departments, the organizing principle for how work gets done.
Imagine smaller, cross-functional teams that own a complete business outcome, such as customer onboarding, from beginning to end. Instead of coordinating work through multiple departmental hierarchies, these teams bring together the necessary human expertise, decision-making authority, and AI agents to execute work across traditional functional boundaries. Accountability shifts from completing departmental tasks to delivering measurable customer and business results.
This doesn't mean eliminating functional departments. They remain important for developing expertise, maintaining standards, allocating specialized capabilities, and providing governance. But functions should enable outcomes, not dictate every step of how work is executed. Smaller outcome-based teams become the primary units of execution, while functional departments provide the capabilities and oversight that support them.
Making this shift requires redesigning workflows, team structures, decision rights for both people and AI, resource allocation, and accountability together.
Automate old workflows and you get faster silos. Restructure teams without redesigning the work and you get smaller silos. Without shared outcome ownership, neither delivers true transformation.
When Incentives Reward Inefficiency
But changing team structures alone won't break silos if the traditional management and incentive systems remain intact.
Consider an outcome-based team responsible for improving customer onboarding. Its success depends on contributions from sales, operations, finance, and technology. Yet the people involved still report to functional leaders whose priorities, budgets, and performance measures are primarily departmental.
When priorities conflict, who decides? The outcome team may own the result, but functional managers still control the resources and decisions needed to achieve it. Even well-intentioned leaders will naturally prioritize the targets against which they're evaluated.
The result is a familiar contradiction: we create cross-functional teams but continue managing them through functional hierarchies.
Traditional productivity measures can reinforce the problem. Employees may be rewarded for the volume of work completed rather than work eliminated through AI, while managers may be recognized for the size of their departments rather than the outcomes they enable.
We cannot expect teams to collaborate around shared outcomes while rewarding their leaders for optimizing individual departments.
Breaking these silos requires more than reorganizing teams. It means aligning functional KPIs with shared business outcomes, giving outcome teams meaningful decision-making authority, and rewarding leaders for collective results rather than departmental optimization. Otherwise, the silos we remove through technology will reappear through management structures and incentives.
Rethinking Human Roles and Leadership
AI is increasingly capable of processing information, executing repetitive tasks, and coordinating activities. Humans remain essential for judgment, creativity, relationships, accountability, and recognizing when established rules no longer make sense. Work must be redesigned around these complementary strengths.
In Workforce: From Jobs to Capabilities, we explored why an employee's value increasingly depends not on the volume of tasks completed, but on the ability to apply judgment, orchestrate intelligent systems, and deliver outcomes. Management must evolve accordingly, moving from supervising activities and collecting updates toward developing people, setting direction, and managing risk.
The deeper transformation, however, starts with leadership.
In Ships and Reefs: The Case for Becoming an Organism, we compared traditional organizations to ships, where intelligence is concentrated at the top and decisions flow downward. The emerging model resembles a living ecosystem, where intelligence is distributed, decisions happen closer to the work, and teams continuously adapt.
This requires a different leadership mindset. Leaders must shift from controlling activities through hierarchical structures to establishing clear direction, empowering outcome-based teams, and defining the guardrails within which people and AI can act. Authority moves closer to where value is created, without sacrificing governance or accountability.
AI capabilities are also evolving faster than many organizations can absorb them. By the time a conventional transformation program finishes, the technology may have advanced again.
The harder challenge isn't simply implementing AI. It's building an organization capable of continuously adapting as AI evolves.
That requires leaders willing to challenge established practices, unlearn what made them successful, and recognize when yesterday's strengths have become today's constraints.
