AI is changing what a normal workday asks of people, and not only by taking tasks away. It adds oversight: checking the output, fixing what the model got wrong, handling the exceptions it can't. IBM's September study of 1,500 CHROs and 8,800 employees found that 80% of CHROs believe AI creates this kind of "invisible" work, and 42% of employees said AI either increases their workload or creates work that goes unrecognized.
We recently looked at how AI adoption became a work-redesign problem that lands on People Ops whether or not anyone hands it over. Buying the tools was the easy part. What comes after the redesign is harder.
Responsibilities are moving, but pay and promotions are not. Payscale found that 55% of organizations had not adjusted pay for the AI skills they now expect. Workday's October 2026 Global Workforce Report adds that internal moves fell to 57% of employers and promotion rates stayed essentially flat worldwide, so a person whose job got bigger has fewer places to take it. Workday's Phil Willburn framed the gap this way: "Employees may not be changing jobs, but their jobs are changing around them."
Pay for what changed in the job
Companies are rewriting jobs faster than they're repricing them, creating a compensation problem.
Employers and employees are already reading the same skill two ways. In Payscale's September preview of its AI Workforce Impact Report, 56% of employees said they should be paid more for developing AI skills, while nearly a quarter of employers now treat AI fluency as baseline. Sometimes fluency is just part of the job. Sometimes it changes what the job is worth. Your job architecture should tell you which.
So the useful question for a Head of People building a compensation strategy is narrower. When AI changed the work, what happened to the value of the job?
Five things to check when AI changes a role
- Tasks. What work disappeared, and what showed up to replace it?
- Skills. Does the role still run on the same skills, or does someone now have to judge AI output?
- Scope. Is this person responsible for more customers, more channels, or a function that used to belong to someone else?
- Accountability. Are they making or signing off on decisions that carry more risk than before?
- Progression. Has the path to the next level changed, especially if the entry-level work people used to climb through is gone?
Don’t focus on how many things moved; focus on which ones did. Changes in tasks tell you how the work changed. Changes in skills might change the role's market value. Scope and accountability tell you whether the job itself got bigger. Progression tells you whether the career architecture still works. When scope or accountability materially change, updating the job description isn't enough. A job is wired into its level, band, performance criteria, and the path above it.
Picture a marketing manager. She runs media campaigns that an AI tool drafts. Every one crosses her desk before it goes live, and she owns the quality of it all. Over the past year, she also picked up managing the blog and social media, two functions that used to belong to other people. Her title is still Marketing Manager, and her salary is still $110k.
AI is what made the consolidation possible, because drafting got cheap enough for one person to direct all three streams. However, ownership didn't get cheaper. Someone still has to answer for what ships, and that someone is her.
Her tasks and skills changed, but those alone describe the same job done faster. Scope and accountability are what make it bigger: she answers for everything the AI ships, and her role now spans work that previously sat across three jobs. That's a re-leveling question. She's doing a bigger job than the one she was hired into, so the review should re-benchmark the role. If a raise follows, it comes out of the re-leveling and has nothing to do with how fluent she is with AI.
Nobody decided to give her a bigger job. It piled up, one handoff at a time, under a title that never changed. Until someone acknowledges this new workload, the pay question can't even be asked.
The version that works in the exec room
Walking into the CFO's office with "employees are worried they aren't getting paid for AI" is a discussion you lose before you sit down. Instead, make the case that AI adoption created a productivity dividend, and the company needs to decide purposefully and strategically where it goes.
Say AI hands a team back 20% of its time. That time can go to more output from the same people, to higher-value work, to fewer people, back to employees as lower load or flexibility, or to work the company couldn't afford before. Any of those is a legitimate call. The failure here is never making a decision, because then the default is "great, you can do more now."
IBM found that only 42% of CHROs say AI's productivity gains are primarily reinvested in innovation or reskilling. Where the rest of that capacity goes is still a decision, whether anyone makes it purposefully or not.
Employees notice which way that goes. Once people see that every efficiency they report turns into more work, they stop reporting it. The 20% is still there, but only they know where it is.
Before that meeting, get straight answers to a few things. Which tasks no longer need meaningful human time? Saying you rolled out Copilot doesn't count. What took its place? More customers, bigger decisions, oversight of the AI, wider scope? Who gets the dividend, and has anyone said so? Are you paying for the job that exists now or the one you hired for? And if the work people used to climb through is gone, what replaced it as a way up?
Where to start
For the next review cycle, pick the two or three functions where AI has absorbed the most real work and the job description now sits furthest from the one actually being done. Run each through the above checklist.
You can expect most to come back as the same job done faster, which then becomes a conversation about output.But a few will come back as bigger jobs, and those should get re-leveled for real: new level, new band, new performance criteria, a career path that makes sense, and a job description to match. Getting the change on the record is what turns absorbed work into recognized work. Then decide with leadership where the freed time goes and tell employees, because if they're left to guess, they'll keep their efficiencies to themselves.
The jobs your employees are doing have already changed, whether your job descriptions reflect it or not. This upcoming review cycle is the chance to separate the roles that simply got faster from the ones that actually got bigger, and make the level, pay, performance expectations, and career path catch up.
