# Agents Aren't Taking Your Jobs. They're Creating More Work Instead.

## Executive summary

AI agents are generating significantly more work for humans—an 'agent management tax'—rather than eliminating it. The complexity of managing these agents scales dramatically from individual use to enterprise deployment. While verifiable domains (like legal or coding) show rapid adoption due to clear success criteria, small businesses often struggle with limited capital and resources. Enterprises gain a significant advantage by having dedicated teams for agent governance, security, and deep integration, which is necessary to manage the increased complexity.

## Key takeaways

- Agents create work, they don't eliminate it: The common assumption that agents will reduce headcount is incorrect. Data shows agent token usage is increasing rapidly (e.g., 14-fold between February and August on Open Router), with agents burning more than five tokens for every one a human burns. This necessitates new management roles.
- The role shifts to 'Above the Loop': As agents improve, the human job is shifting from execution to oversight: deciding what runs, providing context/permissions, checking results, and intervening when failure occurs. This requires domain knowledge (e.g., legal expertise) to validate outcomes.
- Enterprise advantage lies in capital and structure: Enterprises report better returns because they can afford dedicated teams (security, quality control, product management) to handle the complex setup, monitoring, and integration required for agent deployment. This deep investment is necessary for scaling.
- SMBs must focus on verifiable domains: Small businesses struggle when agents are used in non-verifiable domains (e.g., general business operations). Success requires finding processes they already perform manually and letting the agent handle only the preparatory steps.

## Technical details

- Token Usage & Efficiency: Agent token usage is up significantly (e.g., 14-fold between Feb and Aug on Open Router), exceeding human token use by more than five to one. OpenAI reports that heaviest Codex users generate over 60 hours of agent activity daily.
- Verifiable Domains: These are domains where the outcome can be easily proven right or wrong (e.g., law, coding). This verifiability is key to establishing agent management norms and achieving faster adoption.
- Agent Management Tax: The work required for managing agents—including monitoring runs, setting up controls, and disaster recovery—is a new cost/job function that must be factored into business planning.
- Enterprise Integration Requirements: Deep enterprise integration requires granting access across varied tools, adhering to InfoSec policies, and logging all agent runs—tasks that require a dedicated team structure (managerial/security/QC).

## Practical implications

- For SMBs: Focus agent deployment on processes that are already verifiable and where the owner performs manual checks (e.g., reviewing a quote).
- For Vendors/Consultants: Do not sell general 'time savings.' Instead, focus on making inputs easier for the business owner to perform their necessary validation steps.
- For Managers: Recognize that managing agents is becoming a core managerial function ('Agent Allocation Role'), requiring new skillsets in workflow governance and risk assessment.
- Risk Mitigation: Assume all agent runs carry significant risk (e.g., data deletion, as seen with Pocket OS). Disaster planning and strict access controls are mandatory.

## Topics

AI Agents, Workflow Automation, Enterprise Architecture, Digital Transformation, Operational Risk Management, Substack Newsletter, OpenAI Codex Usage Data

Source: https://www.youtube.com/watch?v=IpEaSa7tgfc