From Coding Sidekick to 5 Million-User Knowledge Work Platform
OpenAI’s Codex is an AI agent originally built for coding that has evolved into a broader knowledge worker tool, combining software development assistance with research, reporting, and analytical workflows for both technical and non-technical roles. Codex has now reached 5 million weekly active users, more than six times its level at the desktop app launch in February, showing strong Codex enterprise adoption and consumer interest. Knowledge workers account for about 20 percent of that base and are adopting Codex more than three times as fast as developers, while personal users exceed 5 percent and are growing more than four times as fast. Each week, 72 percent of knowledge worker users generate reports, memos, contracts, images, audio, video, PDFs, and spreadsheets, blurring the line between code work and other AI knowledge worker tools across product management, design, research, and executive tasks.
Knowledge Work Use Cases: Data Analysis, Research, and Parallel Tasks
Codex’s expansion beyond programming is clearest in its task mix. Among knowledge workers, data analysis is the fastest-growing use case, increasing 110 percent week over week, with research up 37 percent and knowledge artifacts up 36 percent. Within those artifacts, work on PDFs and spreadsheets has grown more than 50 percent, turning Codex into a daily spreadsheet analysis and reporting companion. Data labeling accounts for most data analysis volume and growth, while drafting messages, designing products, understanding contracts and regulations, and hiring and interviewing all post more than 40 percent growth. Codex also changes how work is sequenced: about 50 percent of users now keep more than one task running at once, up from below one third in mid-April, orchestrating parallel threads for dataset inspection, script drafting, report assembly, and application checks from a single AI workspace.
Job-Specific AI Plugins and New Collaboration Features
To turn Codex into a practical AI knowledge worker tool for business teams, OpenAI has released six job-specific AI plugins aimed at data analytics, creative production, sales, product design, equity investing, and investment banking. Each plugin bundles tools and context so non-technical users can work “out of the box” on role-specific tasks, aligning Codex enterprise adoption with real workflows. New capabilities deepen collaboration. Sites lets users publish Codex work as hosted interactive websites, from financial scenario planners to product launch hubs, through partners including Wix, Figma, Replit, Base44, Lovable, and Emergent. Annotations, first previewed for developers, now helps knowledge workers direct Codex at specific sections of documents, slides, or spreadsheets for precise edits or analysis. Together, these features shift Codex from a chatbot into a shared workspace where teams create, inspect, refine, and reuse AI-built tools.
AWS and Amazon Bedrock: Making Codex Enterprise-Ready
OpenAI is pairing product features with distribution moves that target IT leaders. Codex and OpenAI’s frontier models, including GPT-5.5, are now available through Amazon Web Services and Amazon Bedrock. This OpenAI AWS deployment lets enterprises access Codex through the infrastructure, security controls, billing systems, and governance frameworks already wired into their cloud operations, easing the shift from pilots to production. Technology teams can tap Codex for code writing, reviewing, debugging, and modernising while also supporting research, reporting, and spreadsheet analysis in the same environment. As organisations focus less on raw model power and more on integration, monitoring, and oversight, Codex’s presence on AWS positions it as a platform that fits into existing compliance and security review processes rather than requiring parallel infrastructure or duplicate governance stacks.
Competing with Claude Code for the Enterprise Knowledge Worker
OpenAI’s Codex push lands in a competitive landscape, especially against Anthropic’s Claude Code and its broader enterprise agents programme. Codex differentiates itself by reaching beyond coding into AI knowledge worker tools that support research, reports, spreadsheet modelling, and job-specific workflows through plugins. According to OpenAI launch materials, “20% of Codex’s 5 million weekly active users are now knowledge workers and not coders,” a signal that the product is becoming an enterprise knowledge worker platform rather than only a developer assistant. Features like Sites, Annotations, and role-aligned plugins for sales, investing, and creative teams aim to lock Codex into daily business operations, while the OpenAI Deployment Company and AWS integrations address concerns around embedding AI into existing infrastructure. The contest with Claude Code is now about who can own the AI-powered knowledge worker desk, not only the IDE.
