OpenAI says its AI tools now cut 40 to 60 minutes from daily work across many jobs. The claim comes from a new company survey run inside real workplaces, not from tests in labs.
The data lands at a tense moment, three years into the AI boom, when many leaders still question what the tech truly gives back.
Workers in data science, engineering, communications, and accounting logged the biggest time cuts. The results focus on task speed, not job loss.
OpenAI tracked how long routine work takes with and without its software inside daily workflows, and allegedly, the main gains came from writing, research checks, light coding, and document tasks.
The survey covered 9,000 employees across 100 companies. All responses came from workers three to four weeks into using the tools on the job. 75% said AI improved either their work speed or the quality of what they deliver.
Heavy users who mixed advanced models and multiple tools saw the largest gains. Casual users saw smaller changes. OpenAI also tracked results by job type and industry to map where time cuts showed up most.
Workers cut task time across key roles
Time savings did not show up the same across every role. Jobs in data science, engineering, communications, and accounting recorded the largest drops in task time.
In those roles, workers said the tools handled drafts, checks, summaries, and basic analysis in minutes instead of nearly an hour. The survey landed while doubts still hang over AI payoff.
In August, researchers at the Massachusetts Institute of Technology said most organizations saw zero return from their generative AI spending. In the next month, teams at Harvard and Stanford said many professionals now push out what they called workslop, meaning AI work that looks clean but lacks depth.
Those studies fed fears of a fresh tech bubble as companies spend billions with no firm payoff path. OpenAI and other AI firms answered with their own reports to show daily business impact.
Last week, rival Anthropic said its Claude model cut task time by 80%, based on 100,000 user conversations. None of those reports went through peer review. The fight over what AI truly returns is now playing out in public numbers.
Enterprises scale paid use at speed
OpenAI now reports more than one million businesses paying for its enterprise AI tools. Paid seats for ChatGPT at work now total seven million users. The company says business adoption now moves as fast as consumer growth, and in some cases even faster. Chief Operating Officer Brad Lightcap spoke on the clash between outside research and company data.
“There’s a lot of studies flying around saying this, that and the other thing,” Brad said. “They never quite line up with what we see in practice.” He added that enterprise use is spreading across teams, not staying locked inside IT.
Usage behavior also shifted. Workers who relied on advanced models and mixed several tools in one workflow logged the strongest gains.
The report also shows people using AI for work they did not touch before. Employees in engineering, IT, and research, who do not hold technical roles, showed a 36% increase in coding‑related messages over the past six months.
Chief economist Ronnie Chatterji tied that shift to new abilities inside offices. “Three out of four people are now saying, ‘I can do things I couldn’t do before,’” Ronnie said. He added that this part of the work shift often gets ignored when AI is debated.
Sharpen your strategy with mentorship + daily ideas - 30 days free access to our trading program
