The position may not have changed, but the deliverables might have already.
In a research tracking study published by OpenAI on September 16, following about 6,200 people who continued using it from April to July: the share of new tasks—ones that previously didn’t belong to their job roles—within AI-related activities rose from 13.1% in April to 25.9% in July. Next month, the proportion of “communicating with customers again” is 54%, ad copy is 44%, and marketing materials is 37%.
This is not a job-replacement rate, nor does it mean you can learn a tool and get a raise. The analysis is based on anonymized, aggregated work messages.
A more useful approach for landing deals and job hunting: pick a cross-role task that’s been repeated more than twice, keep the input, review process, and outcome. On the third time, write it into your portfolio or service list. Hold on to deliverables that can be verified—worth more than a vague “I can use AI.”
Source: OpenAI, “How workers are unlocking new ways of working,” September 16, 2026. Not personally tested.
In a research tracking study published by OpenAI on September 16, following about 6,200 people who continued using it from April to July: the share of new tasks—ones that previously didn’t belong to their job roles—within AI-related activities rose from 13.1% in April to 25.9% in July. Next month, the proportion of “communicating with customers again” is 54%, ad copy is 44%, and marketing materials is 37%.
This is not a job-replacement rate, nor does it mean you can learn a tool and get a raise. The analysis is based on anonymized, aggregated work messages.
A more useful approach for landing deals and job hunting: pick a cross-role task that’s been repeated more than twice, keep the input, review process, and outcome. On the third time, write it into your portfolio or service list. Hold on to deliverables that can be verified—worth more than a vague “I can use AI.”
Source: OpenAI, “How workers are unlocking new ways of working,” September 16, 2026. Not personally tested.