[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.
![[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.](https://substackcdn.com/image/fetch/$s_!roAd!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3a48b472-a807-4211-b981-983f7a401c3a_2156x1576.png)
It's happening.
OpenAI is reporting significant, measurable increases in AI output within internal workflows across various departments, indicating a current inflection point for AI agent adoption and impact.
Sophisticated readers should care because these metrics demonstrate a rapid, real-world productivity multiplier from AI, particularly within knowledge work, impacting labor markets and operational efficiency.
The reported growth suggests that AI agents are moving beyond experimental phases to become deeply embedded and highly productive tools across diverse internal functions, accelerating the automation of tasks.
- · OpenAI
- · Companies adopting AI agents
- · AI developers
- · Knowledge workers leveraging AI
- · SaaS providers for undifferentiated tasks
- · Consulting firms performing rote analysis
- · Legacy workflow software providers
Companies will increasingly invest in and deploy AI agents across their operations to replicate these productivity gains.
This acceleration of AI agent adoption will lead to significant reconfigurations of white-collar workflows and potentially trigger workforce re-skilling or reductions in certain areas.
The demonstrated impact will further fuel the 'AI agents' narrative, potentially leading to increased regulatory scrutiny around job displacement and the ethical implications of autonomous systems.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at Latent Space