The number that should worry higher education
A survey of 1,000 US hiring managers at companies with more than 100 employees, conducted by the career site ResumeTemplates.com and published in late July, found that 48 percent would rather invest in AI tools than hire and train a recent college graduate. That figure represents nearly half of the hiring managers surveyed making an explicit, deliberate choice between a graduate's labor and a software subscription, and choosing the software over the person, in a survey specifically designed to isolate that tradeoff rather than ask about AI adoption in the abstract.
The topline hiring numbers are less alarming on their own: 65 percent of respondents plan to hire the same number or more 2026 graduates compared with last year, and only 23 percent expect to hire fewer or none, with the remaining 12 percent still undecided as of the survey. But the 48 percent figure sits underneath those plans as a structural preference, one that shapes which roles survive the next round of budget planning, how those surviving roles get designed and staffed, and how much genuine training investment a new graduate can expect to receive once they are actually hired into one of them.
Where the money has already moved
This is not a hypothetical preference employers are only weighing for the future. 55 percent of the companies surveyed said they have already reallocated entry-level hiring budget toward AI tools, and 30 percent said AI has directly reduced their overall need for graduate hires. Budget reallocation of that scale, across a majority of surveyed employers, indicates a shift already underway inside hiring plans for the current cycle rather than a theoretical possibility being floated for some future budget year.
The organizational structure emerging from that shift is concrete too. 45 percent of respondents said they now run some functions with a single senior employee paired with AI tools covering work that used to require multiple entry-level hires, and among those companies, 20 percent said a senior worker plus AI is now covering three or more roles that would previously have gone to separate junior staff. That is a meaningful compression of the traditional entry-level rung on the career ladder.
The specific gaps employers describe are not really about AI
Read past the AI headline and the survey's underlying complaints look familiar and largely unrelated to artificial intelligence at all, more a continuation of long standing employer frustrations than a new phenomenon AI has introduced. Nearly 70 percent of hiring managers cited character concerns about recent graduates, with 33 percent specifically pointing to a perceived lack of work ethic among new hires. 76 percent said new graduates need help understanding basic workplace documents such as memos, contracts, and budgets, 41 percent said graduates cannot write a professional email without assistance, and 40 percent flagged weak data analysis and interpretation skills as a recurring problem in their first months on the job.
Only 17 percent of hiring managers said they completely trust new graduates with customer-facing roles, and 61 percent said they would only hire a graduate without prior internship experience for specific limited roles, or would not hire one at all. Taken together, these findings describe a gap in applied workplace fluency, the kind of judgment and polish that internships and workplace exposure build over time, well beyond anything a better AI curriculum alone is designed to close.
Why AI makes this gap more expensive to ignore now
Before AI tools were widely available, the applied skills gap described above was absorbed by employers as a normal, expected cost of hiring anyone straight out of college. A new hire took longer to write a clean email or interpret a budget than a five year veteran would, and companies budgeted training time, manager attention, and a certain amount of patience to close that gap gradually over the first year or two on the job, treating it as the price of building a pipeline rather than a problem to eliminate outright.
AI changes the calculation because it now offers employers a substitute that does not need three months to learn how the company writes internal memos or structures a client email. Once a senior employee can pair with an AI tool to cover the output that used to require two or three separate junior roles, the cost of tolerating a slow ramp-up period for a new graduate rises sharply relative to that alternative, and that recalculation is exactly what shows up in the 55 percent of companies who have already shifted hiring budget away from graduates and toward AI tools instead of training programs.
The pressure this puts back on universities
None of the specific gaps employers cited, understanding a contract, writing a professional email, interpreting a budget, working reliably with a customer, are exotic or new demands. They are exactly the applied workplace skills that career services offices, business communication courses, and internship programs have always been supposed to teach, and the survey suggests that gap has not closed even as AI literacy becomes a more visible part of curricula at many schools.
For university leadership and the edtech vendors serving them, this is a signal to invest specifically in applied, workplace facing skill building rather than assuming AI fluency courses alone answer what employers are asking for. The 61 percent of hiring managers who said they would only consider graduates without internship experience for narrow roles is a particularly direct data point: structured, supervised workplace exposure before graduation is what employers say they are actually screening for.
What this means for the enterprise org chart
For CTOs and operating leaders building out their own teams, this survey is a preview of a staffing model that is likely to keep spreading: fewer traditional entry-level slots, more senior staff paired with AI tools covering broader scopes of work, and a higher bar for the applied skills a graduate needs to demonstrate before getting hired at all. Building that model deliberately, with a clear view of which tasks genuinely require judgment a graduate can develop on the job versus which tasks AI now covers reliably, will matter more than simply cutting headcount and hoping the gap closes itself.
The organizations that get the most value out of this shift will be the ones that treat the compressed entry-level rung as a design problem rather than a budget line to trim. That means being explicit about what a graduate hire is actually being hired to learn on the job, and pairing them with AI tools deliberately rather than assuming either the graduate or the software alone can cover the work that used to take two or three people.
