Robert Half finds 72% of employers pay more for AI skills, even outside tech

AI can automate work, but employers are also paying more for people who know how to use it. Robert Half’s 2027 Salary Guide says 72% of surveyed employers are increasing pay for candidates with relevant AI skills. The premium reaches beyond technology, even as the staffing firm forecasts only modest growth in starting salaries.

A hiring advantage beyond the tech department

The guide’s findings distinguish AI expertise from other technical skills: 42% of employers said it commands a larger pay premium. That is a measure of how many employers favor AI expertise—not the percentage increase a candidate can expect. A separate finding shows competition for specialized talent more broadly: 57% are exceeding planned starting salaries to attract candidates with specialized skills.

Fortune describes the guide as drawing on market data, recruiter insights and surveys of more than 2,000 U.S. hiring managers, business leaders and employed workers. Its evidence therefore combines salary-market information with respondents’ accounts of hiring and compensation, rather than relying on a single source of data.

92%Technology

Technology had the highest reported share among the fields listed in AZFamily’s coverage of the guide.

82%Finance and accounting

More than four-fifths of surveyed finance and accounting leaders said they paid more for AI skills.

60%Administrative and customer support

A majority also reported higher pay for AI skills in administrative and customer support work.

The pattern extends to legal work, where 67% of surveyed leaders reported paying more for AI skills, and marketing and creative work, at 65%. These figures describe the prevalence of a pay premium within each field. They do not compare the dollar value of that premium across occupations.

Starting pay, not an annual raise

The salary forecasts have an important boundary: they concern starting base salaries for new hires, not raises for existing employees. Finance and accounting starting pay is projected to rise an average of 1.8% in 2027, slightly below the guide’s projected national average increase of 1.9%. Several finance roles are expected to grow faster:

• Financial analysts: projected starting-pay growth of 3.8%, compared with the finance and accounting average of 1.8%.
• Staff accountants: projected growth of 3.7% in starting base salaries for new hires.
• Financial planning and analysis managers: projected growth of 3.6%, also above the broader finance forecast.

Outside finance, projected starting-pay growth reaches 3.9% for executive assistants and marketing automation specialists, and 3.3% for paralegals and data scientists. Those are occupation-level forecasts, not estimates of an AI-specific pay boost. Robert Half vice president Travis Laird told AZFamily that companies are bringing executive assistant and paralegal roles back in-house and becoming more selective about hiring.

Location adds another distinction. Robert Half projects Phoenix starting salaries at 13% above national averages. The guide’s national midpoint for a marketing automation specialist is $86,000, compared with about $97,200 in Phoenix. For executive assistants, the corresponding figures are $73,000 and about $82,500. The Phoenix comparison is geographic; it should not be mistaken for the premium attached to an individual candidate’s AI skills.

Using AI—and knowing when it is wrong

Laird’s explanation of the hiring demand is practical: employers are not only seeking people who build AI systems. They also want workers who can automate tasks or use AI features in software they already have. He advises candidates to prepare to explain how they are using AI to work faster or better. Certifications can help, he said, when they match the job.

Finance hiring adds a different requirement. James Tucker, who leads corporate finance and strategy globally at Boston Consulting Group, told Fortune that entry-level hiring is shifting toward judgment rather than task execution. Junior employees increasingly need to evaluate AI-generated work, not simply produce the numbers themselves. His comments describe a hiring shift, not a quantified salary finding. Alongside Laird’s emphasis on using AI efficiently, they identify two distinct capabilities: getting work done with the tools and assessing the work those tools return.