Automation Creates More Expert Work by Redefining Roles

The AI revolution isn't about replacing humans; it's about redefining their roles. This conversation reveals a profound paradox: as artificial intelligence and agents become more capable, the demand for expert human work doesn't diminish--it shifts and, in many ways, intensifies. The hidden consequence isn't mass unemployment, but a new form of overwhelm stemming from an "infinite backlog" of tasks that agents can tackle but humans must still guide and refine. Business leaders, product managers, and anyone navigating the evolving landscape of work should read this to understand how to leverage AI not just for efficiency, but for strategic growth and competitive advantage by investing in human expertise.

The Infinite Backlog: How Automation Creates More Expert Work

The initial promise and fear surrounding AI agents often centered on a future of reduced work hours and increased leisure. However, the reality for early adopters and AI-native companies paints a far more complex picture. Instead of finishing work early, many find themselves compelled to work later, driven by an ever-expanding list of possibilities that agents unlock. This phenomenon, termed the "infinite backlog," signifies a fundamental shift: automation doesn't eliminate work; it redefines its boundaries and elevates the role of human expertise.

At the forefront of this exploration is Dan Shipper, CEO of Every, whose essay "After Automation" dissects this paradox. Shipper argues that AI commoditizes the "residue of human expertise"--tasks that can be made explicit enough to train on. This process collapses the value of default AI outputs, creating a demand for what is unique and different. This difference, he posits, is where human experts thrive, even as we approach more advanced AI capabilities. The common narrative of AI threatening jobs, Shipper contends, overlooks a crucial dynamic: the more we automate, the more complex and nuanced human work emerges.

"The new reality is the opposite: the more we automate, the more expert human work there is to do. Here's why: AI commoditizes the residue of human expertise. Whatever can be made explicit enough to train on, that collapses the value of default model output and creates demand for what's different. And demand for what's different is demand for human experts, even as we approach artificial general intelligence."

This shift manifests in two primary modes of human-agent collaboration. The first, agents as employees, involves delegating tasks to AI entities that operate with a degree of autonomy. These can be "co-worker agents" tagged in Slack for specific tasks, or "embedded agents" within product workflows, handling repetitive functions. While these agents automate stable, repeatable work, they still require human oversight for complex decision-making and framing.

The second, and arguably more significant, mode is human-agent collaboration, exemplified by tools like Codex and Claude Code. These are not merely delegation platforms but "operating systems for work itself." Here, humans and multiple agents work concurrently on highly complex, original tasks that asynchronous agents cannot fully execute. This collaborative environment, described as a "human sandwich," involves humans setting the goals and defining success, AI collapsing tasks into drafts and summaries, and humans then judging, extending, and directing the work. This model acknowledges that for complex tasks, the synergy between AI and human judgment is paramount.

"For the last several weeks, AI has responded to 95% of my work emails. In short, the future looks weird, but also familiar. The familiarity is surprising because one thing CEOs, knowledge workers, and investors seem to agree on is that AI is a threat to jobs, the economy, safety, and human meaning."

The evolution of agent deployment also highlights a move away from individual agents towards shared, team-based agents. Initially, companies experimented with giving each employee a personal AI agent. This approach proved burdensome, as individuals were responsible for maintaining and fixing their own agents. The realization dawned that team-based agents, where multiple individuals rely on a single, shared agent for specific functions (like analytics), offer greater efficiency, easier maintenance, and continuity. This shift from personal assistants to shared co-workers underscores the systemic benefits of well-managed AI resources.

The AI discourse often predicts a tipping point where AI replaces human workers entirely. However, the evidence suggests a more nuanced outcome. AI's exponential trajectory, trained on past human competence, commoditizes existing skills. This abundance of "cheap competence" leads to sameness in default AI outputs, paradoxically creating a demand for human experts who can provide differentiation and address emergent needs. Humans, unlike AI, possess an understanding of what needs to be done now, driving innovation and requiring expert judgment. This dynamic suggests that AI doesn't eliminate expert human knowledge work; it expands the volume of work and necessitates human involvement for value creation.

"Abundance creates sameness. Old expertise becomes commoditized because everyone has access to the same models, and the models are all based on yesterday's competence. By default, the models end up creating work that ranges from a decent start to it's just plain slop... But sameness creates a demand for difference. Humans, Dan points out, very quickly spot this sameness and want something better. Demand for difference, Dan argues, is new demand for experts."

Ultimately, the most successful companies in the AI era will be those that invest in their teams' capabilities to manage and leverage agents, recognizing that growth, not just efficiency, is the key to long-term success. Agents are not a cost-cutting measure but a significant investment opportunity, fostering AI-related growth and expanding the horizons of what humans and AI can achieve together.

Key Action Items

• Invest in Shared Agent Infrastructure: Over the next quarter, identify overlapping work areas within your teams and explore deploying shared, team-based agents for common tasks (e.g., analytics, content summarization). This reduces individual maintenance burden and enhances team synchronicity.
• Develop "Human Sandwich" Workflows: Immediately begin mapping out workflows where human oversight frames the task, AI executes the core work, and humans refine and direct the output. Focus on complex, original tasks where this model offers the greatest advantage.
• Prioritize Agent Management Training: Within the next six months, implement training programs focused on managing AI agents effectively. This includes understanding their limitations, guiding their execution, and interpreting their outputs, rather than simply delegating.
• Shift Focus from Efficiency to Growth: This is a continuous, long-term investment. Reorient strategic planning to prioritize AI-driven growth opportunities rather than solely focusing on cost reduction. Recognize agents as a catalyst for new business models and market expansion.
• Experiment with Semi-Synchronous Collaboration: Over the next 1-3 months, encourage individuals and teams to experiment with tools that reduce latency between human instruction and AI execution. Focus on compressed feedback loops for more fluid, real-time collaboration.
• Cultivate Demand for Differentiation: This is an ongoing cultural shift. Foster an environment where identifying and creating unique value, distinct from commoditized AI outputs, is a core competency. Reward human expertise that provides novel insights and solutions.
• Reframe AI Investment: View AI agents not as a "get out of budget jail free" card, but as a strategic investment with significant long-term payoff potential. This requires a shift in financial planning and ROI assessment, looking beyond immediate cost savings to future value creation.