I think HR should give one named person the job of checking internal candidates before any AI-related requisition opens. The people who answer to that job today are split on whether it is theirs, and a new survey shows what that split costs.
The numbers behind the argument
Careerminds, a workforce-solutions provider, published a white paper on November 4 called The AI Redeployment Gap, built on a Censuswide survey of 1,142 U.S. HR professionals and people managers at organizations with 500 or more employees. Of the HR professionals, 92% said employees in their organization could be reskilled for AI-related roles, and 88% of people managers agreed. Yet 51% of HR professionals and 47% of managers said their organizations are hiring externally for new AI positions, against 42% in each group who said they are redeploying existing staff.
The release points to a visibility problem. About 43% of HR professionals and 34% of managers reported limited visibility or uncertainty about which employees should be reskilled, and 87% of HR professionals still said their organizations adequately consider reskilling for future roles. The survey ran from May 15 to 26, 2026.
Ownership shows the sharpest split. Of the HR respondents, 63% said HR holds primary responsibility for planning and managing AI-driven workforce change. Only 24% of managers agreed, and 35% of managers pointed to senior leadership outside HR.
Why I read this as an ownership problem
When two-thirds of one group believe they own a decision and a quarter of the other group agrees, the decision is made somewhere else, usually by whoever holds the requisition. A manager with budget and a deadline posts the role. Nobody is positioned to ask whether a person already on payroll could do it in ninety days with support.
Mark Saddic, Senior Partner at Careerminds, put the stakes in the release: “Before looking outside the organization, leaders need to understand the skills they already have, identify where employees can grow, and establish a clear process for making those transitions possible.” That describes a process, and a process needs an owner with authority to pause a posting. A survey finding that 87% of HR professionals think reskilling is adequately considered, set beside a 42% redeployment rate, suggests the consideration happens in conversation and not in the workflow.
The white paper, as the release describes it, cites separate research from The Josh Bersin Company indicating that external hiring can cost three to five times more than filling comparable roles internally and take up to 20 days longer. I would treat that as a reason to test the claim against your own cost-per-hire and time-to-fill, since the release does not say how those comparisons were built.
The strongest objection
The best argument against my position is that some AI roles need skills no current employee has, and that waiting on an internal search delays work the business has already funded. That is true for some roles. The release itself notes that the gap varies by industry: in Technology and Telecom, external hiring outpaces redeployment by 15 percentage points among both groups, while Media and Entertainment is the one sector where both groups report redeployment at least as often as external hiring. A firm in the first group may have good reasons to hire from outside more often.
The rule I am proposing allows for that. It asks for a documented check, time-boxed to a few days, with the result recorded on the requisition. If the check finds nobody, the posting proceeds and HR has a record of why. If it finds a candidate, the company has saved the search. The cost of the rule is a short delay on each role, and the survey suggests many organizations cannot currently say how many of their AI hires could have been internal moves.
What the labor data adds
Challenger, Gray & Christmas reported this week that U.S. employers announced 43,994 job cuts in October. Andy Challenger said, “Employers are holding on to the workers they have heading into the holidays.” Our report on the October numbers also covers the 60,000 voluntary exit offers Microsoft announced, which Challenger says tend to signal more layoffs to come. Companies shrinking by attrition and buyouts while opening AI requisitions are releasing people with institutional knowledge and hiring others to learn it again.
I made a related argument when the data showed that HR is analyzing AI’s workforce impact but not replanning around it. A redeployment owner is the smallest concrete step toward replanning, because the role forces the question of who in the current workforce maps to which future job.
What I would do
Name the owner this quarter, and give that person three things: read access to skills data, a seat in requisition approval for AI-related roles, and a short written standard for what counts as a qualified internal candidate. Report quarterly how many AI-related roles were filled internally and externally, so the 51% and 42% become your own numbers. Start by reading yesterday’s piece on writing role-level AI expectations before building a skills library, since an internal check needs those expectations to compare against.
