Every HR technology vendor is promising leaders that AI agents are about to remake how work gets done. Almost none of them are promising that the managers now expected to run that transformation are ready for it, and the data piling up this month says they are not.
Three separate pieces of research published in the past two weeks, from a management consultancy, an HCM vendor, and Gallup’s workplace institute, converge on the same structural problem: organizations are redesigning work faster than they are preparing the people who have to manage it day to day. The shift is not that managers have gotten worse at their jobs. It is that the job itself has expanded, often overnight, while the support underneath it has not moved at the same speed.
The Readiness Gap Is Now Measurable
Deloitte’s “Path to Agentic Transformation” report, based on a survey of 501 senior managers and C-suite executives across five industries between April and June 2026, found that fewer than half of leaders say their organization is ready for agentic AI across most aspects of the business. Only 16% say their business processes are actually prepared for agentic adoption today, even though 74% expect at least half of those processes to be redesigned around AI agents within four years. Seventy-two percent of leaders cite the lack of a unified data foundation as a barrier, and 70% say trust and governance of autonomous agents remains unresolved.
That gap between ambition and readiness lands hardest on the management layer. Workforce-transformation consultancy Paradigm’s “State of Workforce Transformation 2026” report, drawn from benchmark data across roughly 700 organizations, found that fewer than 10% of organizations feel most of their managers are adequately prepared to lead through the kind of change now underway. The report frames this as a widening “middle manager squeeze”: expanding spans of control and accelerating change, layered onto managers who were never trained to run either.
The Numbers Behind the Squeeze
Gallup’s workplace research puts a hard figure on that expansion. The average span of control, the number of people reporting to a single manager, rose from 10.9 in 2024 to 12.1 in 2025, a nearly 50% increase since 2013. Gallup’s analysis, covering more than 92,000 teams across 104 organizations in 26 industries, also found that managers now spend a median of 40% of their time on individual-contributor work rather than managing, meaning the job has grown wider at the same time it has been squeezed for capacity.
HCM vendor Dayforce put similar pressure into a different lens. Its research, surveying nearly 5,700 respondents across Australia, Canada, Germany, New Zealand, the U.K. and the U.S., found that 77% of front-line managers say their organization’s systems lack the guidance they need when operational issues come up, and 67% say decisions slow down because workforce data is scattered across disconnected platforms, numbers this publication covered in detail last week. Only 6% of executives and managers say transformation initiatives actually integrate into daily work; 80% say the daily grind of operations disrupts or outright prevents transformation from landing.
“When workforce data and workflows are disconnected, managers spend too much time searching for answers and not enough time acting on them,” said Steve Holdridge, President and Chief Operating Officer at Dayforce. “The organizations best positioned for what comes next won’t simply launch more initiatives.”
How We Got Here
The logic behind wider spans of control is not new, but it has accelerated. Flatter organizational charts have been a cost-cutting staple for years; what changed in 2026 is the argument for why a flatter chart is now supposedly safe. AI tools that automate scheduling, reporting and performance monitoring, the administrative core of a manager’s job, are being used to justify handing each remaining manager a bigger team, on the theory that software absorbs the paperwork a wider span used to require. Gallup’s data complicates that theory rather than confirming it: median team size has stayed flat at five to six employees even as the average has climbed toward 12, meaning the increase is concentrated in a smaller group of managers being pushed well past what the research shows most teams can absorb well.
Gallup’s own analysis identifies four factors that determine whether a wider span works or breaks a team: how engaged the team already is, how much of the manager’s time goes to individual-contributor work rather than managing, the manager’s own talent and capability, and the quality and frequency of the feedback that manager gives. Its research found that highly engaged teams of 12 or more, double the current median of six, can thrive under an effective manager, while poorly managed teams struggle even at a fraction of that size. In other words, span of control is not the variable that predicts outcomes on its own. Manager capacity, and whether the organization has actually checked that capacity before widening it, is.
Why This Is a Systems Problem, Not a Training Problem
The instinct inside most HR functions has been to answer a readiness gap with more training: a workshop, a coaching program, a resource library, the same instinct that left HR’s own confidence in manager readiness running well ahead of managers’ self-reported confidence. Paradigm’s research suggests that instinct is aimed at the wrong layer. Two-thirds of organizations in its sample already provide development resources to managers, and half connect them with coaching support. Despite that investment, the preparedness numbers have not moved. What is missing, according to the report, is not content but visibility: only 10% of organizations systematically analyze span-of-control data to see where the gaps actually are before assigning more responsibility to a manager who is already stretched.
Deloitte’s findings point at the same disconnect from the leadership side. Its report describes organizations racing to redesign processes around AI agents without first fixing the data foundation and governance structure that would make redesign safe, let alone effective. One healthcare enterprise AI architect quoted anonymously in the report put it bluntly: redesign, done properly, is currently “too expensive” for most organizations to attempt at the pace leadership wants, which pushes the burden of managing an unredesigned process onto whoever is closest to the work, the frontline manager.
The workforce-impact numbers in the same Deloitte survey show why that burden is not temporary. Forty-three percent of leaders anticipate “a lot” to “extreme” disruption to jobs within the next 12 to 18 months, rising to 72% who expect significant disruption within two to three years. Seventy-one percent say baseline AI literacy efforts are already underway and 65% are running targeted reskilling programs, yet half of leaders surveyed say their own organization still is not investing enough in workforce transformation. That is a leadership population that largely agrees the disruption is real and immediate, and is still not funding the preparation for it at the pace it says the disruption requires. The mismatch between what leaders say they expect and what they say they are funding is, in effect, the same readiness gap Paradigm found in managers, one level up the org chart.
What It Means for the HR Leader
Read together, the four data sets describe a single mechanism rather than four separate trends. Leadership sets a transformation timeline based on what the technology can theoretically do. Spans of control widen because flatter structures and AI-assisted administration make it look affordable to give one manager more people. Data and workflow systems stay fragmented because integration is expensive and rarely funded ahead of the reorganization it is meant to support. The manager in the middle absorbs all three pressures at once, with a fraction of the guidance the job now requires.
For HR and people-ops leaders, that reframes the readiness question. It is not “how do we train managers for AI transformation.” It is “what decision-support and data visibility does a manager need before we widen their span of control or hand them an AI-redesigned process.” Dayforce’s data suggests the fastest-moving lever is not a new course but a fix to how workforce data reaches the manager in the first place: centralizing scheduling, performance and operational data so a manager is not searching six systems to answer one question. Gallup’s research adds a second lever: measuring span of control explicitly, team by team, rather than assuming a flatter chart is automatically a more efficient one.
What to Do Before the Next Reorganization
Three moves show up consistently across the research. First, audit span of control and manager workload before finalizing any flattening decision, using the kind of systematic analysis Paradigm found only 10% of organizations currently do. Second, treat data and workflow integration as a prerequisite for AI-driven process redesign, not a cleanup task for afterward; Deloitte’s numbers show organizations that skip this step are the same ones reporting the widest gap between ambition and process readiness. Third, build the feedback loop that tells HR when a manager’s span has crossed from stretched to unsupportable, rather than waiting for turnover or a disengagement survey to surface it months later.
None of this requires abandoning agentic AI or flatter structures; the pressure toward both is not going away. What the data argues against is treating either as free. A manager handed more people, a less defined process and an unintegrated set of systems is not being empowered. They are being asked to absorb a redesign that was never actually finished, and every one of these reports says that absorption is already showing up in how managers describe their own jobs.
The clearest signal that this is solvable, rather than an inevitable cost of AI-driven restructuring, is Gallup’s finding that wide spans and strong engagement are not mutually exclusive when a manager has the capacity to support them. That reframes the HR leader’s job in the next reorganization cycle away from picking a target ratio and toward diagnosing, team by team, whether the manager taking on more people also has the data, the feedback loop and the time back from administrative work that would let a wider span succeed instead of quietly failing. The organizations named across all four reports that are furthest ahead are not the ones moving fastest on flattening the chart. They are the ones that checked, before they flattened it, whether the managers left standing could actually run what remained.
Source: Deloitte
