From Coding Sidekick to General-Purpose Knowledge Work Platform
Codex is an AI productivity tool that began as an agentic coding assistant and has evolved into a general-purpose platform for knowledge workers, supporting research, analysis, content creation, and interactive apps in a single workspace. OpenAI reports that Codex now has 5 million weekly active users, more than six times its level when the desktop app launched in February, and that roughly 20 percent of these users are knowledge workers instead of developers. These Codex knowledge workers create reports, memos, contracts, images, audio, video, PDFs, and spreadsheets every week, blurring the line between software tasks and traditional office work. Product managers now build dashboards on their own, researchers write dataset-cleaning scripts, designers ship prototypes without a dedicated developer, and executives assemble internal tools for reconciled reporting. Codex has moved from being a specialist tool for engineers to a shared workspace for entire teams.
Sites and Annotations: Turning AI Output into Live Workspaces
The Sites feature signals how AI for business analysis is shifting from chat transcripts to living tools. Sites lets teams create interactive dashboards, websites, and lightweight apps, all shareable by URL inside a workspace. OpenAI describes this as moving beyond “a chatbot that answers questions” toward a place where teams bring in role-specific context, create work, refine it in place, and then turn it into something colleagues can use. A financial model can become a scenario planner; a launch checklist can become a site with live milestones and messaging. Annotations deepen this shift. Users can select a passage in a document, a range in a spreadsheet, or a section of a slide deck and ask Codex to edit or use that specific slice as context. Instead of copying content between tools, Codex plugins features keep the AI’s actions anchored to the exact artifact already in progress.
Role-Specific Codex Plugins Bring AI to Every Desk
OpenAI’s latest Codex plugins features are aimed squarely at non-developer teams, turning Codex into an AI hub for cross-functional work. Six role-specific plugins connect Codex to data analytics, creative production, sales, product design, public equity investing, and investment banking workflows. Together, they span 62 apps and 110 skills, plugging into tools that analysts and business teams already rely on, including platforms for data warehouses, notebooks, and dashboards. For data analysts, Codex can now help with reports, dashboards, and complex data labeling tasks, which the sources describe as the largest and fastest-growing part of data analysis usage. Education providers and workforce skills teams can use the same tool to train staff on AI for business analysis, from dashboards and research outputs to creative briefs and internal apps. Codex is turning into a shared toolkit for analysts, marketers, operators, and investors, not only software engineers.
Canva and Creative Production: Design Work Inside Codex
Creative production shows how Codex knowledge workers now span far beyond engineering or analytics roles. The creative production plugin connects Codex to tools such as Figma, Canva, Shutterstock, Picsart, and Fal, so a marketing brief drafted in Codex can become design-ready assets without leaving the workspace. Canva’s GM and Head of Ecosystem, Anwar Haneef, noted that “Canva is now inside Codex” and described this as “the next step” in the companies’ collaboration since DevDay 2023. He emphasized that when an idea takes shape in Codex, it can move straight into Canva “on-brand, fully editable, ready to share.” The design layer follows the work instead of forcing teams to manage exports and uploads between tools. For marketers and designers, Codex is becoming one of the central AI productivity tools that ties copy, visuals, and approvals together in one environment.
Parallel Workstreams and the Future of AI-First Organizations
Codex is also reshaping how individuals structure their day. About 50 percent of users now keep more than one parallel task running in Codex, up from less than one third in mid-April. A single person can inspect a dataset in one thread, write a script in another, assemble a report in a third, and check an application in a fourth, effectively orchestrating multiple AI workstreams. Among knowledge workers, data analysis tasks are growing 110 percent week over week, with research up 37 percent and knowledge artifacts up 36 percent. Use of PDFs and spreadsheets has grown more than 50 percent, underscoring how often Codex is used as AI for business analysis and documentation. As plugins and features like Sites and Annotations spread across departments, Codex is moving from a specialist coding tool to a shared platform for organization-wide knowledge work.
