OpenAI says Codex has reached 5 million weekly active users, with usage in Australia rising sharply since January.
The figures suggest broader adoption beyond software development. Half of Codex use in Australia now comes from non-technical tasks, according to OpenAI, with marketers, analysts, designers, researchers, investors and bankers among those using the tool.
OpenAI also introduced role-specific plugins and tools for teams using Codex in routine office work. The additions are intended to help users automate workflows, create dashboards, build presentations, generate creative assets and turn analysis into shareable applications.
Its report on knowledge work says Codex is increasingly being used for reports, spreadsheets, presentations, contracts, research, data analysis and workflow automation. Developers remain the largest user group, but knowledge workers now account for about 20% of users and are growing more than three times as fast, the report says.
In Australia, Codex use has risen sixfold since January and 13-fold among enterprises, OpenAI says. That suggests some of the fastest growth is coming from businesses rather than individual technical users.
Beyond Coding
The shift marks a change in how OpenAI is positioning Codex. Once seen mainly as a coding assistant, it is now being presented as a broader workplace tool for non-technical staff handling tasks that often relied on engineering teams or specialist software users.
The fastest-growing knowledge-worker tasks are data analysis, research and knowledge artifact creation, according to the report. OpenAI says users are increasingly running several tasks at once, allowing them to investigate data, draft materials and automate workflows in parallel.
That pattern supports the report's broader argument that modern office work has become fragmented across email, chat platforms, documents, dashboards, databases and ticketing systems. OpenAI says this fragmentation has led to delays in finding information, coordinating across teams and getting approvals.
The report cites data showing knowledge workers spend about 28% of their time managing email and nearly 20% searching for information or colleagues who can help. It argues that these frictions have become a major drag on productivity even as digital tools have become more common.
Role Expansion
OpenAI says AI tools are beginning to blur role boundaries inside organisations. In examples from the report, product managers create dashboards, researchers write scripts, designers build prototypes and executives create internal tools without waiting for engineering support.
It argues this lets staff closest to a problem act directly rather than pass work through multiple teams. In turn, that could reduce the time spent moving information between departments and waiting for reviews or specialist input.
Half of Codex users now run multiple tasks simultaneously, up from less than one-third in mid-April, according to the report. OpenAI describes that as a sign that users are beginning to manage several AI-driven workstreams at the same time.
The report places the trend in the context of the wider economy, noting that knowledge work now accounts for more than 40% of the US workforce, or about 72 million people. It says the burden of administrative and coordination tasks has grown as work has spread across more digital systems.
Australian Uptake
The Australian figures stand out because they indicate strong adoption outside technical teams. With half of local usage tied to non-technical tasks, the market offers an example of how AI tools once associated mainly with programmers are moving into mainstream office functions.
That may be especially relevant for larger employers, given the reported 13-fold rise in enterprise usage since January. OpenAI's new plugins and tools appear aimed at those organisations, particularly teams looking to automate recurring work without building bespoke software.
The report also calls for AI literacy to become a core workforce skill and for public-sector procurement to focus more directly on operational problems. It says worker-led adoption and updated public workflows would help organisations make practical use of AI systems in day-to-day work.
Overall, OpenAI says the trend shows Codex is being used to reduce the friction of modern work by helping people find information, coordinate tasks across systems, produce deliverables and move projects through review and approval processes.