Codex use is spreading into knowledge work, OpenAI says

2026-07-01 · Source: Dataconomy · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Intermediate, quick

Summary

OpenAI's recent research paper, "The Shift to Agentic AI: Evidence from Codex," reveals its Codex platform has become the dominant AI tool internally, generating 99.8% of output tokens among employees by June 11, 2026. All departments, including Legal and Finance, now utilize Codex for work-related tasks, with non-technical teams reaching majority usage by April 2026. Legal roles saw a 13-fold increase in monthly output tokens, and researchers a 50-fold increase since November 2025. External non-developer usage surged 137 times for individual users and 189 times for organizational users since August 2025. The report notes a significant shift from conversational to agentic AI, where users assign multi-step tasks. Task complexity is rising, with 25.6% of individual users submitting requests estimated to require over eight hours of human work, up from 2.1% in December 2025. Codex now boasts over 5 million weekly active users.

Key takeaway

For AI Product Managers evaluating enterprise AI solutions, OpenAI's internal Codex adoption highlights the critical shift towards agentic AI for complex tasks. You should prioritize tools that enable multi-step workflows and measure productivity gains across all departments, not just technical ones. Be prepared for users to assign increasingly complex work to AI, requiring robust and scalable solutions.

Key insights

Codex's internal adoption demonstrates a significant shift towards agentic AI for complex, multi-step knowledge work.

Principles

• Internal AI adoption accelerates rapidly under favorable conditions.
• Agentic AI drives substantial productivity gains across diverse roles.
• User-submitted task complexity on AI platforms is increasing.

In practice

• Deploy AI for multi-step knowledge work tasks.
• Track employee output token growth with AI tools.
• Monitor the increasing complexity of AI-assisted tasks.

Topics

• Codex
• Agentic AI
• OpenAI
• Enterprise AI Adoption
• Knowledge Work Automation
• Productivity Gains
• Task Complexity

Best for: Executive, Investor, Entrepreneur, Director of AI/ML, AI Product Manager, Consultant

Related on AIssential

• OpenAI's Codex update lets agents build interactive enterprise workspaces via Sites and role-specific plugins — Codex's update transforms it into an agentic AI operating environment for business, enabling precise automation and interactive app creation.
• [AINews] OpenAI reports median internal Codex output tokens grew 56x in Research, 32x in Customer Support, 27x in Engineering, and 13x in Legal since November 2025. — Rapid internal AI agent adoption at OpenAI, alongside advancements in open models and agent infrastructure, signals a significant shift in AI application.
• OpenAI Codex Knowledge Work Expansion: What Changed and When — OpenAI expanded Codex from a coding assistant to a general knowledge-work platform, targeting diverse professions beyond its initial 5 million developer users.
• OpenAI Codex App: death of the VSCode fork, multitasking worktrees, Skills Automations — Agent-native coding interfaces and specialized open models are rapidly advancing, shifting developer workflows and evaluation paradigms.
• OpenAI expands Codex with enterprise plug-ins and new Sites feature — OpenAI is rapidly expanding Codex's enterprise utility and user base beyond developers with new features and strategic partnerships.
• My Adventures Building an AI Money Machine — OpenAI's Codex enables non-technical users to build functional AI systems for file-based "real work" beyond chatbot capabilities.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Dataconomy.