AI moved what a team can produce. In most places it moved nothing about what the job is, who reviews the work, or what a full week looks like.
In 89% of organizations, fewer than half of roles have been updated to reflect what AI can now do. That's from Workday's survey of 3,200 leaders and employees, who describe it as people running current tools inside job structures written years ago1.
The capability moved. The container it sits in didn't.
When output speeds up and the role stays fixed, the extra work has to go somewhere. It goes into the hours nobody assigned.
ActivTrak's Productivity Lab tracked 443 million hours of activity across 1,111 companies. Time spent working on Saturdays rose 46%. Sundays rose 58%. People also started their Saturdays an hour and twenty-four minutes earlier than they used to. Meanwhile the weekday itself shrank by about 2%2.
Put plainly, the weekday got slightly shorter and the weekend got busier. The week didn't grow where anyone measures it. It grew on Saturday morning.
Nobody asked for Saturday. The expectations moved faster than the structure around them, and conscientious people closed the gap out of their own time because nobody told them not to.
Buying the tool is a decision most teams made deliberately. These three tend to go unmade.
| Moved With The Tool | Stayed Where It Was |
|---|---|
| How fast a first draft appears | The job description |
| How much output gets expected | The definition of a full week |
| The number of tools in the workflow | The review process |
| What counts as a rough first pass | Who gets coached, and on what |
One more number from the same research. Two thirds of leaders name skills training as a top investment priority. Only 37% of the employees dealing with the most rework say they're getting more access to it1.
Leaders mean it. The training is just planned against the old role. When the job description says one thing and the daily work has become another, training aimed at the documented job keeps missing the people doing the undocumented one.
Take one role on your team and read its actual description. Not the recruiting version, the one in the system.
Then mark every line that AI has changed in the last year. Faster, easier, or no longer the point. Most people mark between a third and half the document.
Rewrite three lines with the person who does the job. Not the whole document, three lines. The conversation surfaces the expectations nobody had written down. That's the point of the exercise, more than the rewrite itself.
AI made the work faster. It didn't update the role, the review process, or the standard for a full week, and it was never going to. Those are written by people, and in most organizations they haven't been rewritten yet.
Until they are, your team will keep closing the gap themselves, on Saturdays, and you'll find out about it from a resignation rather than a report.
Up Next: The Job That Changed the Most
One role has shifted further than any other, and it isn't the one doing the selling.
Revenue operations used to be a reporting and systems function. Now the same team is being asked to decide how work gets redesigned around AI, which tools are allowed, and who owns a workflow once software sits in the middle of it. That's change management, and most people in the role haven't been trained for it. The next post covers what the job has turned into and what it needs.
Sources
Numbered citations throughout this article draw on the sources above.
