AI agents create more work, not less: meet the Cowan Paradox

The Gist

• SaaStr debunks the 'do more with less' AI hype, showing how automation actually raises customer expectations and workload.
• Their team now runs 20+ AI agents alongside 3 humans to maintain $1M revenue streams.

• AI agents don't reduce work; they reset the baseline of what's expected, leading to more work.
• The Cowan Paradox explains that labor-saving technology raises standards and creates new categories of work.
• AI adoption leads to task expansion, multitasking, and cognitive overload, intensifying work rather than reducing it.
• Companies deploying AI agents must focus on leveraging AI to do previously impossible tasks, not just replicate existing workflows.
• AI equalizes productivity across companies, making differentiation dependent on human judgment, insight, and creativity.
• The cognitive load and volume of decisions increase as AI handles execution, requiring intentional norms to prevent burnout.

• Focus on leveraging AI to accomplish tasks that were previously impossible, not just replicating existing workflows.
• Implement intentional norms and guardrails to manage cognitive load and prevent burnout as AI intensifies work.
• Differentiate through human judgment, insight, and creativity, as AI equalizes productivity across competitors.
• Redeploy human resources to high-leverage work while AI handles the expanded baseline of expected output.

• 51 hours/week (Time spent on housework by housewives from the 1930s to 1950s, unchanged despite labor-saving appliances.)
• -19% to +47% (SaaStr's revenue trajectory year-over-year after deploying AI agents.)
• 500,000+ users (SaaStr.ai reached this milestone in 45 days.)
• 750,000+ uses (Number of uses generated by AI applications built by the author.)
• 1 million customer problems/week (Resolved by Intercom's AI agent Fin.)
• 140 meetings/7 days (Booked by Personio's AI chat agent.)