Agentic Work Adoption

Agentic work adoption is the measurable shift from using AI as a conversational assistant to using agents such as Codex to delegate, monitor, and review multi-step work.

Key points

• OpenAI's public article frames Codex as a work-delegation interface, not only a coding assistant ^[src-169].
• The underlying paper separates individual users, organizational users, and OpenAI workers so adoption is not treated as one blended number ^[src-170].
• OpenAI workers are the frontier case: by June 11, 2026, Codex accounted for 99.8% of output tokens across Codex and ChatGPT for that population ^[src-170].
• Organizational users show a weaker but still meaningful shift, with Codex at 63.3% of combined output tokens versus 16.5% among individual users ^[src-170].
• The practical adoption markers are concurrency, skills, task complexity, and runtime, not only whether a user has access to an AI tool ^[src-170].
• OpenAI's internal usage should be treated as directional: the paper says OpenAI is unusually favorable because employees have strong AI familiarity, buy-in, cheap marginal usage, and frequent knowledge sharing ^[src-170].

Related entities

• Openai
• Codex

Related concepts

• Everyday Agentic Work
• Agentic Workflows
• Human Agent Collaboration
• AI Productivity Measurement
• Agent Skills
• Agentic Workplace Collaboration

Source references

• ^[src-169] OpenAI – "How agents are transforming work" (2026-06-25)
• ^[src-170] Drew Johnston, David Holtz, Alex Martin Richmond, Christopher Ong, Prasanna Tambe, Aaron Chatterji / OpenAI – "The shift to agentic AI: Evidence from Codex" (2026-06-25)