The Two-Job Job: Role Sprawl, Hidden Workload, and How Leaders Can Redesign Work Before It Breaks People
Download the article here:
Listen to a review of this article:
Abstract: Delayering, hiring restraint, unbackfilled attrition, and generative AI deployment have combined to produce what recent coverage calls the two-job job: roles that quietly absorb the responsibilities of more than one position without being redesigned. Drawing on Korn Ferry's Workforce 2026 survey of more than 16,000 professionals, in which 61% reported doing more than one role and 62% reported significantly heavier workloads, this article examines role sprawl as a work design problem rather than a matter of individual resilience. It integrates role theory, the job demands-resources model, burnout research, work design scholarship, and emerging evidence on the hidden labor created by organization-wide AI initiatives. The article traces organizational costs, including the activity-productivity gap, stalled innovation, and turnover, and individual costs such as exhaustion, declining motivation, and inequitable distribution of invisible work. It then sets out five evidence-based responses, illustrated with examples from technology, finance, pharmaceuticals, healthcare, and professional services, and concludes with three long-term capabilities for governing workload deliberately.
Ask almost any mid-career professional to describe their job in 2026 and listen to how often the answer contains the word "also." The analytics lead who also covers vendor management after a colleague left without being replaced. The clinic manager who has also become the department's informal AI champion. The senior engineer who also absorbed the people management of a team whose manager role was eliminated. These are composites rather than reported cases, but the pattern they describe is now well documented. None of these roles was formally redesigned or upgraded. Each simply grew sideways, one absorbed task at a time, until it became two jobs wearing the name of one.
That pattern now has a number attached to it. In its Workforce 2026 Global Insights Report, a survey of more than 16,000 professionals across 11 markets, Korn Ferry (2026) found that 61% of respondents said they were carrying the responsibilities of more than one role, and 62% said their workloads had become significantly heavier over the past two years. Fortune's coverage gave the phenomenon a memorable label, the rise of the "two-job job" (Lin, 2026). The same survey suggests the extra load is not translating into extra value: 45% of respondents said they were too busy to deliver meaningful results that contribute to growth, and reported motivation had fallen from 71% in 2024 to 61% in 2026 (Korn Ferry, 2026).
What makes this moment distinctive is the convergence of forces that each add work while subtracting capacity. Organizations have spent several years removing management layers and holding headcount flat. Attrition goes unbackfilled. Generative AI tools are being deployed with promises of efficiency, yet building, testing, and governing those tools generates its own stream of labor that rarely shows up in anyone's job description (Karunakaran et al., 2026). Earlier pieces in this journal examined the flattening of management structures and the paradox of AI time savings that turn into intensified work. This article takes a different angle. It focuses on the unit where all of these pressures finally land: the individual role.
My argument is straightforward. Role sprawl is a work design problem, and it requires a work design response. Telling people to prioritize, offering resilience workshops, or rolling out another productivity tool treats the symptom while leaving the structural cause intact. Decades of research on role theory, job demands and resources, and work design give leaders a usable toolkit for diagnosing and correcting overloaded roles. A growing set of organizational experiments, from calendar purges to formal programs for removing unnecessary tasks, shows what that toolkit looks like in practice. The sections that follow define the core concepts, examine the drivers and consequences of role sprawl, and lay out five evidence-based responses before turning to the longer-term capabilities organizations need so that roles can flex without breaking the people in them.
The Role Sprawl Landscape
Defining Role Sprawl, Role Overload, and Hidden Work
The vocabulary matters here, because several related ideas tend to get blurred together in management conversation.
Role overload is the oldest and best-established of these concepts. In their foundational study of organizational stress, Kahn and colleagues (1964) described roles as bundles of expectations that other people, called role senders, place on a focal person. Overload occurs when the combined expectations exceed what a person can reasonably accomplish in the time available. Kahn et al. (1964) also distinguished overload from role conflict, where expectations are incompatible with one another, and role ambiguity, where expectations are unclear. Rizzo, House, and Lirtzman (1970) later developed widely used measures of role conflict and ambiguity and showed that both were associated with lower satisfaction and higher anxiety. These distinctions are practically useful. A person doing two jobs often experiences all three at once: too much to do, competing demands from two sets of stakeholders, and fuzzy boundaries about which responsibilities are truly theirs.
Role sprawl, as I use the term, describes the structural process by which a role accumulates responsibilities beyond its original design without a corresponding redesign of the role itself. Sprawl is distinct from job enlargement, which is a deliberate decision to broaden a role, and from promotion, which typically comes with new authority, pay, and support. Sprawl happens by accretion. A colleague leaves and their tasks are "temporarily" distributed. A manager position is eliminated and coordination work drifts to the most senior individual contributor. A new system is introduced and someone becomes its unofficial administrator. Each increment seems small and reasonable; the sum is a role nobody designed.
Hidden work, sometimes called invisible work, refers to effort that is necessary for outcomes but is not formally recognized, measured, or rewarded. Star and Strauss (1999) showed that what counts as visible work is socially negotiated, and that much of the articulation work that keeps systems running (coordinating, troubleshooting, translating between groups) goes unseen by those who evaluate it. Two contemporary forms of hidden work are especially relevant. The first is collaborative work. Cross, Rebele, and Grant (2016) reported that time spent in collaborative activities had grown by 50% or more over two decades and that, across more than 300 organizations, a small group of 3% to 5% of employees supplied 20% to 35% of value-added collaborations. The second is the work generated by AI adoption itself, which I discuss in detail below.
Finally, the job demands-resources (JD-R) model offers the most useful integrating framework. Bakker and Demerouti (2007, 2017) proposed that every job can be described in terms of demands, the aspects of work that require sustained effort and carry physiological or psychological costs, and resources, the aspects that help people achieve goals, reduce the cost of demands, or support growth. High demands without matching resources drive a health-impairment process toward exhaustion; ample resources drive a motivational process toward engagement. Role sprawl is, in JD-R terms, a mechanism that steadily raises demands while frequently stripping resources such as managerial support, clear goals, and autonomy over priorities.
Prevalence, Drivers, and Distribution
The Korn Ferry (2026) survey provides the clearest current picture of prevalence. Beyond the 61% performing more than one role and the 62% reporting significantly heavier workloads, 44% said they felt stretched beyond their capabilities. The management picture is equally striking. Korn Ferry (2026) reported that 42% of organizations had cut management roles in the past year, that 55% of the managers who remained described themselves as exhausted, and that 39% of workers said a lack of managers left them feeling directionless.
These figures come from a single, large, multi-country survey released in a press format, and they should be read as indicative of direction and magnitude rather than as precise population estimates. Still, they are consistent with a broader pattern visible in both corporate announcements and independent workplace data. Four drivers stand out.
The first driver is delayering. A wave of high-profile restructurings has explicitly targeted management layers. In March 2023, Meta announced it would lay off 10,000 people and close 5,000 open roles as part of an effort that Mark Zuckerberg had earlier described as flattening the organization and removing some layers of middle management (Fischer, 2023). Citigroup's chief executive, Jane Fraser, announced in November 2023 a reorganization that would cut the bank's management layers from 13 to eight, alongside the elimination of 60 committees (Reuters, 2023). In September 2024, Amazon's Andy Jassy set a goal of increasing the ratio of individual contributors to managers by at least 15% by the end of the first quarter of 2025 (Lake, 2024). Bayer, under chief executive Bill Anderson, adopted a model it calls Dynamic Shared Ownership, cutting roughly 5,500 jobs, most of them managerial, and asking teams to self-organize around 90-day cycles (Hogg, 2024). Most recently, Uber said in September 2026 that it would cut roughly 3,300 roles, including about 20% of managerial positions (Lin, 2026; Schaal, 2026). The work those managers performed (coordination, coaching, escalation, prioritization) does not disappear when the role does. Some of it is genuinely redundant. Much of it migrates.
The second driver is headcount restraint and unbackfilled attrition. Hiring freezes and "do more with less" budgets create sprawl quietly, through vacancies rather than announcements. The work of a departed colleague is distributed among peers on an interim basis that often becomes permanent.
The third driver is AI deployment. The Korn Ferry (2026) report noted that 52% of what it called AI-weary workers said AI had increased their workloads, and it described managing AI alongside regular tasks as something that can feel like doing two jobs. HR Dive's coverage noted that Korn Ferry researchers linked the workload problem to leaner workforces in which headcount reductions have been largely fueled by AI (Colvin, 2026). The perception gap is notable too: 79% of CEOs reported improved efficiencies from AI, compared with 51% of individual contributors (Korn Ferry, 2026). Leaders who see efficiency at the top may be making staffing decisions based on gains that are not being felt at the front line.
The fourth driver is the steady growth of coordination overhead. Asana's Anatomy of Work Index 2021, based on a survey of more than 13,000 knowledge workers in seven countries, found that respondents spent 60% of their time on what the company calls "work about work," such as searching for information, chasing status updates, and switching between tools, and that 7 in 10 had experienced burnout at least once in the preceding year (Asana, 2021). Microsoft's analysis of anonymized Microsoft 365 telemetry found that the average employee was interrupted roughly every two minutes by a meeting, email, or notification, received 117 emails and 153 Teams messages on a typical weekday, and that meetings starting after 8 p.m. had risen 16% year over year (Microsoft, 2025). These are vendor studies with commercial interests in the problems they describe, but they converge on a pattern: the coordinating layer of work has expanded even as formal coordinators have been removed.
Role sprawl is not evenly distributed. Several groups appear to absorb more than their share. Managers who survive delayering inherit wider spans and more direct reports, which helps explain Korn Ferry's finding that more than half of remaining managers report exhaustion. High performers and helpful colleagues attract disproportionate requests, a pattern Cross et al. (2016) documented in their network research, where the most sought-after collaborators often became bottlenecks and burned out. There is also a gendered dimension. Babcock, Recalde, Vesterlund, and Weingart (2017) found in field and laboratory studies that women were asked more often than men to take on tasks with low promotability, and accepted those requests at higher rates, largely because of a shared expectation that women would say yes. Bolino and Turnley (2005) found that higher levels of individual initiative, the willingness to go beyond formal requirements, were associated with greater role overload, job stress, and work-family conflict, with the link to work-family conflict stronger for women. When orphaned tasks are redistributed informally, they tend to flow toward the conscientious, the visible, and the people least likely to refuse.
Organizational and Individual Consequences
Organizational Performance Impacts
The intuitive case for absorbing work into existing roles is economic. If three people can do the work of four, the organization saves a salary. The evidence suggests that this arithmetic is incomplete, because overloaded roles generate costs that rarely appear in the same budget line as the savings.
The first cost is the activity-productivity gap. The most telling figure in the Korn Ferry data may be the 45% of respondents who said they were too busy to deliver meaningful results that contribute to growth (Korn Ferry, 2026). HR Dive summarized the report's message as a warning that high activity does not mean high productivity (Colvin, 2026). People doing two jobs triage. They protect the urgent and the visible, and they defer the important-but-not-urgent work (process improvement, customer insight, developing others, experimenting with new approaches) that drives future growth. In a classic ethnography of software engineers, Perlow (1999) described a "time famine" in which constant interruptions and crisis-driven work crowded out the focused effort needed for the most valuable tasks. Role sprawl institutionalizes that famine.
The second cost operates through the stressor pathway. LePine, Podsakoff, and LePine (2005) conducted a meta-analysis testing a framework that distinguishes challenge stressors, demands that people tend to experience as opportunities for growth and achievement, from hindrance stressors, demands experienced as obstacles. They found that hindrance stressors had negative direct effects on performance and negative indirect effects through strain and reduced motivation. Role sprawl is a potent generator of hindrance demands: ambiguous ownership, conflicting priorities from different stakeholders, and procedural friction inherited from roles nobody redesigned. Even the parts of an enlarged role that might feel like stretch assignments become hindrances when they arrive without authority or resources.
The third cost is withdrawal and turnover. Bowling, Alarcon, Bragg, and Hartman (2015) meta-analyzed 336 studies on workload and found that heavier workload was associated with poorer psychological and physical well-being, lower affective commitment, and higher turnover intentions and absenteeism. The individual correlations were modest, but the direction was consistent, and the same analysis found that workload was substantially correlated with role conflict and work-to-family conflict. In organizational terms, the people most likely to leave an overloaded system are often those with the most options, which tends to mean the most capable. Each departure, in turn, feeds the next round of redistribution.
The fourth cost is stalled innovation, and this is where AI enters the picture most sharply. Karunakaran, Kellogg, and Wiesenfeld (2026) studied generative AI innovation over two years at an academic medical center and a corporate law firm. They argue that the largest returns from generative AI are likely to come from organization-wide solutions rather than off-the-shelf personal productivity tools, and that building such solutions depends on domain experts committing time to experimenting with the technology and redesigning their workflows around it. That collective experimentation, however, adds work on top of already demanding jobs, a hidden cost that can derail ambitious AI initiatives if leaders do not manage it (Karunakaran et al., 2026). Fortune's account of the research reported that this work included trial-and-error experimentation, reviewing and refining outputs with colleagues in other departments, and continually adapting solutions as the underlying models changed, and that much of it was invisible to leadership (Lin, 2026). Organizations that cut capacity in anticipation of AI gains may thereby starve the very process that would produce those gains.
The fifth cost is weakened direction and coordination. When Korn Ferry (2026) found that 39% of workers felt directionless for lack of managers, it pointed to a function that delayering often underestimates. Managers translate strategy into priorities, broker resources, and protect their teams from conflicting demands. When those functions disappear, role senders multiply and conflict rises, which is exactly the condition Kahn et al. (1964) associated with strain and reduced effectiveness.
Individual Wellbeing and Stakeholder Impacts
For individuals, the core risk is burnout. Maslach and Leiter (2016) describe burnout as a psychological syndrome arising from chronic interpersonal stressors on the job, characterized by overwhelming exhaustion, cynicism and detachment, and a sense of ineffectiveness. They identify workload as one of six areas of worklife in which a mismatch between people and their jobs predicts burnout, alongside control, reward, community, fairness, and values. Role sprawl tends to create mismatches in several of these areas simultaneously. Workload rises. Control falls, because the person no longer chooses what to drop. Reward lags, because the extra role is rarely compensated. Fairness erodes when people see that redistribution landed on them and not on others.
The JD-R model explains why this combination is so corrosive. Bakker and Demerouti (2017) describe a loss cycle in which strain leads to self-undermining behaviors, such as making mistakes, communicating poorly, or creating conflict, which in turn raise demands further. A person doing two jobs who starts missing deadlines generates follow-ups, rework, and escalations. Those additions land on an already overloaded role and accelerate the slide.
Motivation is the most immediate casualty. Korn Ferry's reported decline in motivation, from 71% to 61% in two years, is shown alongside the CEO-employee perception gap on AI. Both data points point to the same disconnect. Leaders see a leaner, more technologically enabled organization; employees experience a heavier, more fragmented job. Lesley Uren of Korn Ferry put the point directly, observing that cost-cutting and restructuring create savings but not the passion that powers productivity (Korn Ferry, 2026).
The effects spill well beyond the individual. Bowling et al. (2015) found that workload was more strongly related to work-to-family conflict than to family-to-work conflict, a reminder that overloaded roles are paid for at home. Customers and patients absorb the consequences as well. In healthcare, documentation and administrative demands compete directly with patient contact, which is why health systems have invested so heavily in attempts to release clinical time for care (Ashton, 2018; Sarre et al., 2019). In professional services, the experts whose judgment clients pay for are the same experts asked to build and validate new AI tools on top of billable work (Karunakaran et al., 2026).
There is also an equity dimension that leaders should not overlook. If redistributed work flows disproportionately toward women and toward highly conscientious employees, as the evidence on non-promotable tasks and individual initiative suggests (Babcock et al., 2017; Bolino & Turnley, 2005), then role sprawl can quietly widen gaps in advancement. The people carrying the invisible second job are doing work that keeps the organization functioning, but it is often work that performance systems do not see and promotion committees do not reward.
Evidence-Based Organizational Responses
The good news is that role overload is one of the most thoroughly studied conditions in organizational psychology, and work design research offers practical levers. Parker, Morgeson, and Johns (2017), reviewing a century of work design scholarship, argue that the way tasks, activities, relationships, and responsibilities are organized shapes outcomes for individuals, teams, and organizations, and that work design is something organizations actively choose, whether or not they do so deliberately. Role sprawl is what happens when that choice is left to drift. The five responses below represent deliberate alternatives. Table 1 summarizes them.
Table 1: Evidence-based organizational responses to role sprawl
Response | Evidence Base | Key Practices | Illustrative Examples |
Subtract before you add | Work design; hindrance stressors (LePine et al., 2005); Ashton (2018) | Work inventories, removal sprints, meeting resets, bureaucracy reporting channels | Shopify calendar purge; Hawaii Pacific Health; Amazon bureaucracy mailbox |
Re-charter roles after restructuring | Role theory (Kahn et al., 1964; Rizzo et al., 1970); Morgeson & Humphrey (2006) | Role charters, explicit decision rights, orphaned-task audits, transition plans | Citigroup layer reduction; Bayer Dynamic Shared Ownership; Meta and Uber delayering |
Govern workload as a capacity budget | JD-R theory (Bakker & Demerouti, 2017); Bowling et al. (2015) | Capacity planning, work-in-progress limits, protected focus time, trade-off rules | Dropbox core collaboration hours; Microsoft telemetry insights |
Make hidden work visible and count it | Star & Strauss (1999); Cross et al. (2016); Babcock et al. (2017) | Collaboration mapping, AI build time in roles, reward assists, rotate low-promotability tasks | Academic medical center vs. law firm AI study; network analysis practices |
Resource managers and enable job crafting | Wrzesniewski & Dutton (2001); Tims et al. (2012); Korn Ferry (2026) | Sustainable spans, manager capacity, crafting conversations, experimentation support | Bayer self-organizing teams; Korn Ferry manager findings |
Subtract Before You Add
The most underused move in work design is removal. Organizations are generally skilled at adding initiatives, reporting requirements, approvals, meetings, and tools, and remarkably poor at taking them away. Every addition has a sponsor; no one owns subtraction. Yet the evidence on hindrance stressors suggests that removing obstacles may be one of the most efficient ways to improve both wellbeing and performance (LePine et al., 2005). The Asana (2021) finding that a majority of knowledge-worker time goes to coordination and administration, rather than to the skilled work people were hired for, implies that a meaningful share of any overloaded role consists of tasks that could be simplified, automated, consolidated, or dropped entirely.
The logic is especially important under role sprawl. When two jobs collapse into one, the natural assumption is that the person must now do the full content of both. A better starting assumption is that neither job was perfectly designed in the first place, and that the merger is an opportunity to strip out the low-value elements of each before asking anyone to carry the remainder.
Effective approaches:
Work inventories: Before redistributing a departing colleague's work, list every recurring task and decide explicitly whether to keep, simplify, automate, reassign, or stop it.
Removal sprints: Run time-boxed campaigns that invite employees to nominate tasks, reports, and requirements that add little value, with a visible commitment to act on nominations.
Meeting resets: Periodically clear recurring meetings and require owners to justify re-adding them, with clear purposes and decision rights.
Bureaucracy reporting channels: Give employees a direct line to report unnecessary process, and publish what changed as a result.
Stop-doing lists at the team level: Pair every new priority with an explicit decision about what will stop or slow down to make room.
Shopify offers one of the most widely publicized examples of a deliberate subtraction. In January 2023, the company announced that it was removing all recurring meetings involving more than two people from employees' calendars, banning meetings on Wednesdays, and restricting large meetings of more than 50 people to a single window on Thursdays (Bove, 2023). A company spokesperson said the purge eliminated roughly 10,000 calendar events, the equivalent of more than 76,500 hours of meetings, and employees were asked not to add meetings back for at least two weeks (Bove, 2023). Chief operating officer Kaz Nejatian framed the change as useful subtraction. The approach was intentionally blunt and, by the company's own description, chaotic; its value lay less in the specific rule than in forcing every team to rebuild its coordination from a blank slate and justify each meeting it restored.
Hawaii Pacific Health took a more targeted approach in a setting where hidden documentation work had been eroding clinical time. Beginning in October 2017, the health system invited clinicians to nominate electronic health record requirements they considered poorly designed or unnecessary, under a program candidly titled Getting Rid of Stupid Stuff (Ashton, 2018). Within the initial period it received 188 nominations from nurses and physicians, and it completed changes on dozens of them. One change removed a required documentation row that nurses had to complete during patient rounding; Ashton (2018) estimated that this single click had been consuming about 1,700 nursing hours per month across the system's four hospitals. The health system also removed 10 of the 12 most frequently triggered physician alerts, which clinicians had largely been ignoring. Ashton (2018) was careful to note that it was too early to measure effects on engagement, but the program demonstrated that substantial capacity can be recovered by asking the people doing the work what should be removed.
Amazon built a version of this channel into its 2024 delayering effort. Alongside the goal of raising the ratio of individual contributors to managers, Andy Jassy established a mailbox through which employees could report unnecessary bureaucracy; Fortune reported that he had received more than 500 submissions and that the company had acted on about 150 of them (Lake, 2024). The pairing is instructive. Removing managers without removing the processes those managers administered simply moves the processes onto someone else's desk.
Common pitfall: Subtraction campaigns often decay. The NHS Productive Ward program, designed to release nurses' time for direct patient care, was widely adopted in English acute hospitals around 2008 and 2009, yet a ten-year evaluation found little robust evidence of a sustained increase in direct care time, and found that financial and management support had disappeared in most trusts (Sarre et al., 2019). Removal works only if it becomes a recurring discipline with an owner, rather than a one-time cleanup.
Re-Charter Roles After Restructuring
The second response addresses role ambiguity and conflict directly. When restructuring changes who does what, the formal and informal expectations attached to each role need to be renegotiated explicitly. Kahn et al. (1964) showed that strain rises when role senders hold conflicting expectations and when the focal person lacks clarity about what is expected. Rizzo et al. (1970) found that both role conflict and role ambiguity were associated with negative outcomes. Morgeson and Humphrey (2006), in developing the Work Design Questionnaire, demonstrated that jobs can be described across a broad set of task, knowledge, social, and contextual characteristics, and that these characteristics relate meaningfully to satisfaction, motivation, and other outcomes. A redesigned role is therefore more than a longer task list. It includes decision authority, interdependence with others, required knowledge, and the support available to do the work.
Restructuring announcements typically describe what is being removed (layers, positions, committees) far more precisely than what each remaining role is now responsible for. That gap is where sprawl breeds.
Effective approaches:
Role charters: For each significantly changed role, document purpose, core outcomes, the responsibilities that were added, those that were removed, and the decision rights that come with them.
Orphaned-task audits: After any elimination of positions, identify every recurring task the departed roles performed and assign an explicit owner or an explicit decision to stop.
Decision-rights mapping: Clarify who decides, who is consulted, and who is informed for the key decisions previously handled by eliminated managers.
Stakeholder alignment: Bring the multiple role senders for an enlarged role together to agree on priorities, rather than leaving the role holder to arbitrate between them.
Interim-to-permanent reviews: Set a date by which any "temporary" coverage arrangement must be either formalized with appropriate adjustments or ended.
Citigroup's restructuring illustrates both the opportunity and the complexity. The bank's reduction from 13 management layers to eight was accompanied, according to its earnings materials, by the elimination of 60 committees and a reduction of functional roles in the top two leadership layers (Reuters, 2023). Eliminating committees is a form of subtraction that can genuinely remove work. The design challenge is ensuring that the decisions those committees made, and the coordination they provided, are deliberately reassigned rather than informally absorbed by whoever is left.
Bayer offers a more radical version. Its Dynamic Shared Ownership model removed large numbers of managerial positions and replaced annual budgeting with self-directed teams working in 90-day cycles, with resources reallocated each quarter (Hogg, 2024). Chief executive Bill Anderson reported a year into the change that early-adopting teams were performing well, that other groups had been slower to start, and that voluntary attrition had fallen even as layoffs rose (Hogg, 2024). Whatever its eventual results, the model is explicit about something most delayering efforts leave implicit: if managers are removed, the work of setting priorities and allocating resources must be redesigned into a new mechanism, not simply left to emerge.
The contrast with less structured delayering is instructive. Meta's 2023 flattening and Uber's 2026 cuts to managerial roles were both framed around speed and simplicity (Fischer, 2023; Schaal, 2026). Those goals are legitimate. The risk lies in assuming that fewer layers automatically mean less work, when much of what managers did must still be done by someone.
Common pitfall: Re-chartering only the roles that were formally changed. In practice, sprawl spreads laterally to peers and adjacent teams who quietly pick up the slack. A charter process that reviews only the roles named in the restructuring plan will miss much of the absorbed work.
Govern Workload as a Capacity Budget
The third response treats work capacity the way organizations treat money: as a finite resource that must be budgeted, with explicit trade-offs when new demands arrive. Most organizations have rigorous processes for approving headcount and spending, and almost none for approving claims on people's time. A new initiative, report, or system rollout can be launched without anyone asking whose hours it will consume and what those people will stop doing.
The JD-R model provides the logic for this response. If strain results from demands that persistently exceed resources, then managing the ratio requires visibility into both sides (Bakker & Demerouti, 2017). The meta-analytic evidence that workload is associated with poorer well-being, lower commitment, and higher turnover intentions suggests that unmanaged demand has real costs (Bowling et al., 2015). Capacity governance also connects to the collaborative overload research. Cross et al. (2016) recommended mapping the supply of and demand for collaboration across a network to identify who is at risk of overload, and setting norms such as requiring an owner and agenda for every recurring meeting.
Effective approaches:
Capacity planning for initiatives: Require that every significant new initiative identify the people whose time it will consume and the existing work that will be deprioritized.
Work-in-progress limits: Cap the number of concurrent projects a team or individual can carry, so that adding new work forces a visible trade-off.
Protected focus time: Establish organization-wide windows for synchronous collaboration and protect the rest of the day for concentrated work.
Trade-off rules: Give managers and employees an agreed process for escalating when demands exceed capacity, with a commitment that escalation leads to reprioritization rather than exhortation.
Telemetry with care: Use aggregated, anonymized data on meetings, after-hours activity, and messaging volume to spot overload patterns at team level, without individual surveillance.
Dropbox designed capacity rules into its Virtual First model. According to a Future Forum case study, the company established four-hour core collaboration windows, set by time zone, for synchronous work, and adopted asynchronous work as the default for solving problems (Future Forum, n.d.). It also offered a simple test for meetings: a gathering should exist to discuss, debate, or decide, and without one of those purposes the meeting may be unnecessary (Future Forum, n.d.). The value of such rules lies in making the default protective of capacity, so that each claim on shared time has to justify itself.
Microsoft's own research on what it calls the infinite workday illustrates how telemetry can surface overload patterns that individuals cannot see. Its analysis found that nearly half of employees described their work as chaotic and fragmented, that one in three said the pace of work over the past five years made it impossible to keep up, and that ad hoc calls without calendar invitations made up a large share of meetings (Microsoft, 2025). Microsoft recommended concentrating effort on the small share of work that drives most outcomes and redesigning teams around goals. Data of this kind, aggregated to the team level, give leaders an evidence base for capacity decisions that otherwise rely on anecdote.
Common pitfall: Governing the calendar instead of the workload. Meeting-free days and focus blocks help, but if the total volume of assigned work does not change, people simply move the overflow to evenings. Microsoft's finding that late-evening meetings had grown is a warning that protected time can be undermined by unprotected demand (Microsoft, 2025).
Make Hidden Work Visible and Count It
The fourth response tackles the invisible second job directly. Work that is not seen cannot be planned for, resourced, or rewarded. Star and Strauss (1999) argued that the visibility of work is shaped by organizational choices about what to measure and recognize. The research on collaborative overload and non-promotable tasks shows that these choices have predictable distributional consequences, concentrating unrecognized work on helpful, conscientious, and often female employees (Babcock et al., 2017; Cross et al., 2016).
AI adoption has created a new and fast-growing category of hidden work. Building organization-wide AI solutions requires domain experts to experiment, evaluate outputs, coordinate with colleagues across functions, and adapt tools as models change (Karunakaran et al., 2026). If that work is treated as an informal extra on top of an existing job, it competes directly with the work people are measured on, and it tends to lose.
Effective approaches:
Count AI build work as real work: Allocate explicit time for experimentation, evaluation, and tool maintenance in the roles of people asked to build or validate AI solutions.
Reward the work in formal systems: Incorporate contributions to shared tools, collaboration, and innovation into performance reviews and promotion criteria.
Map collaboration demand: Use network analysis or simple request tracking to identify people carrying disproportionate collaborative load.
Rotate low-promotability tasks: Assign necessary but unrewarded tasks through transparent rotation rather than volunteering.
Track assists: Recognize contributions that help others succeed, alongside individual output.
The research by Karunakaran et al. (2026) provides the clearest current illustration of what is at stake. According to Fortune's account of their two-year study, at the corporate law firm more than 80% of the domain experts involved eventually dropped out of AI innovation efforts, and the firm ended up with three organization-wide AI tools (Lin, 2026). The academic medical center, by contrast, built AI work into job responsibilities, performance reviews, and promotions, and had developed 141 tools (Lin, 2026). The two organizations differ in many ways, and a comparison of two field sites cannot isolate a single cause. The contrast is nonetheless consistent with the researchers' broader argument: without the right scaffolding, the people closest to the work tend to quietly withdraw from innovation, while formally recognizing the work makes it sustainable (Karunakaran et al., 2026).
Cross et al. (2016) describe a set of practices drawn from their work with organizations on collaborative overload. These include mapping who receives the most requests, steering helpers toward the forms of collaboration that are most valuable and energizing, asking requesters to use existing resources before seeking hands-on help, and factoring collaboration into reviews, promotions, and pay. They report that in one Fortune 500 technology company, 60% of employees wanted to spend less time responding to ad hoc requests. The practical lesson is that hidden work can be measured with reasonable effort, and that measurement is the precondition for redistribution.
Common pitfall: Making hidden work visible without changing anything. Surveys and dashboards that document overload but are not followed by changes in role design, staffing, or recognition can deepen cynicism. Visibility is a means, not an end.
Resource Managers and Enable Job Crafting
The fifth response operates at two levels: strengthening the managers who remain, and giving employees structured latitude to reshape their own roles. Both are resources in JD-R terms, and both buffer the effect of demands on strain (Bakker & Demerouti, 2017).
On the manager side, the Korn Ferry (2026) data are sobering. More than half of the remaining managers reported exhaustion, and 39% of workers said a lack of managers left them directionless. Yet the same survey found that 52% of workers credited their managers with encouraging experimentation, and Korn Ferry recommended that managers receive more organizational support to help teams innovate and test new ways of working (Korn Ferry, 2026). Managers are not overhead to be minimized; they are a capacity-allocating resource. When spans widen, the work managers are expected to do must narrow accordingly, or their own roles become the most extreme case of sprawl.
On the employee side, job crafting research offers a complementary tool. Wrzesniewski and Dutton (2001) defined job crafting as the physical and cognitive changes individuals make to the task or relational boundaries of their work, and argued that employees actively shape their jobs rather than passively receiving them. Tims, Bakker, and Derks (2012) operationalized crafting within the JD-R framework, distinguishing efforts to increase structural and social job resources, increase challenging demands, and decrease hindering demands. Under role sprawl, the decreasing-hindrances dimension deserves attention: employees often know which absorbed tasks are low-value but lack permission to drop them.
Effective approaches:
Sustainable spans: Set span-of-control targets in light of the complexity of the work and the experience of team members, and reduce managers' administrative load as spans grow.
Manager capacity support: Provide shared services, tooling, and peer support for the coordination tasks that formerly required additional managerial layers.
Structured crafting conversations: Hold regular discussions in which employees and managers review the current shape of a role and agree on what to drop, simplify, or reshape.
Permission to decline: Establish clear norms for when employees can say no to requests outside their role, and back those norms with leadership support.
Experimentation support: Protect time and resources for teams to test new ways of working, including with AI.
Bayer's Dynamic Shared Ownership model, discussed earlier, is relevant here as well. By shifting resource allocation to teams working in 90-day cycles, the company effectively attempted to move some prioritization authority from managers to the people doing the work (Hogg, 2024). That shift is a form of organization-wide job crafting, and its success depends on whether teams actually receive the authority to stop work, not just to start it.
Common pitfall: Treating job crafting as an individual coping strategy in an unchanged system. Crafting is valuable when it operates within a deliberate redesign. When employees are left to craft their way out of structural overload alone, the most capable negotiators shed work onto the least powerful colleagues, and the total burden does not change.
Building Long-Term Work Design Capability
The responses above can relieve immediate overload. Preventing sprawl from recurring requires capability: the organizational habits, structures, and skills that keep roles deliberately designed as conditions change. Three capabilities matter most. Role sprawl tends to compound when capacity is removed without redesign, and where these levers can interrupt the cycle.
Workload as a Governed Organizational Resource
The first long-term capability is institutionalizing workload governance. In most organizations today, headcount is governed with great precision while the workload placed on that headcount is ungoverned. Every restructuring business case calculates the salaries saved; few calculate the hours redistributed or the work that will stop. Closing that gap means treating capacity as a governed resource with the same seriousness as budget.
In practice, this involves several changes. Restructuring proposals should include a work disposition plan that specifies, for every eliminated role, which work will be stopped, automated, consolidated, or reassigned, and to whom. Major initiatives, including AI deployments, should include a capacity impact statement that estimates the time required from domain experts during design, testing, and maintenance (Karunakaran et al., 2026). Regular workload health indicators, drawn from pulse surveys and aggregated collaboration data, should be reviewed at leadership level alongside financial and operational metrics. The Korn Ferry (2026) finding that CEOs and individual contributors perceive AI efficiency gains very differently suggests that leaders cannot rely on their own vantage point to judge whether capacity is adequate.
Governance also requires accountability for subtraction. Someone needs to own the question of what the organization will stop doing. The decay of the NHS Productive Ward program after support was withdrawn (Sarre et al., 2019) shows that removal efforts without durable ownership fade. A standing process, whether a quarterly review of reports and meetings, an annual audit of approval requirements, or a permanent bureaucracy reporting channel with published outcomes (Lake, 2024), keeps subtraction from becoming a one-time event.
Role Architecture That Flexes Without Sprawling
The second capability concerns how organizations define roles in the first place. Traditional job descriptions are static documents, written for hiring and rarely revisited. Modern work requires roles that can flex as priorities shift, but flexibility without structure is precisely how sprawl happens. The challenge is building role architecture that can adapt deliberately.
Work design research suggests some principles. Hackman and Oldham's (1976) job characteristics model identified skill variety, task identity, task significance, autonomy, and feedback as core characteristics associated with motivation and satisfaction. Sprawled roles often score poorly on task identity, because the person is doing fragments of several jobs rather than a coherent whole, and on autonomy, because their priorities are set by multiple competing stakeholders. Morgeson and Humphrey (2006) broadened the picture to include social characteristics such as interdependence and support, and contextual characteristics such as work conditions. A well-architected role preserves a coherent core of meaningful work, clear ownership of outcomes, and adequate support, even as peripheral responsibilities change.
Practical applications include defining roles by core outcomes and a limited number of flex responsibilities, with explicit rules for how flex responsibilities are added and removed. Organizations can maintain a role register that records when responsibilities are added to a role and triggers a review when the additions exceed a threshold. Compensation and leveling systems should recognize sustained increases in scope, so that people carrying two jobs are either rewarded for the larger role or relieved of part of it. Parker et al. (2017) emphasize that work design shapes not only immediate well-being but also learning and development over time. Roles designed with that in mind treat added responsibilities as growth only when they come with the resources and recognition that make growth possible.
Manager and Employee Capability for Ongoing Job Redesign
The third capability is human: the skills and norms that allow managers and employees to redesign work continuously, rather than waiting for a formal restructuring. Most managers have never been trained to design work. They are taught to set goals, give feedback, and manage performance, but rarely to analyze a role, identify hindrance demands, or negotiate trade-offs with stakeholders. Building that capability may be the highest-leverage investment an organization can make against sprawl.
Manager development should include practical work design skills: conducting work inventories, facilitating crafting conversations, mapping collaboration demands, and escalating capacity constraints constructively. Managers also need permission and support to push back on demands that exceed their team's capacity. The Korn Ferry (2026) recommendation that managers receive more organizational support recognizes that managers cannot protect their teams from overload if they are themselves overloaded and unsupported.
For employees, the relevant capabilities include recognizing and naming sprawl, negotiating role boundaries, and crafting their work in ways that reduce hindrances without shifting burden onto colleagues (Tims et al., 2012; Wrzesniewski & Dutton, 2001). Organizational norms matter as much as individual skill. Where saying no is treated as a lack of commitment, employees will absorb work until they break. Where clear role boundaries are respected and trade-off conversations are normal, sprawl becomes visible early, while it is still easy to correct.
These capabilities reinforce one another. Governance creates the data and the decision forums; role architecture creates the structure; manager and employee capability creates the everyday practice of redesign. Together they shift work design from an episodic activity, triggered by crisis or restructuring, to a continuous organizational competence.
Conclusion
The two-job job is not an accident of individual time management. It is the predictable product of organizations removing capacity faster than they remove work, while simultaneously introducing technologies whose implementation requires substantial human effort. The Korn Ferry (2026) data capture the scale of the problem: most professionals report doing more than one role, workloads have grown sharply, motivation has fallen, and the managers who remain are exhausted. Decades of research on role overload, job demands and resources, burnout, and work design explain why this matters and point toward solutions.
The central lesson is that role sprawl requires a design response. Organizations that treat overload as a problem of individual resilience will continue to lose capable people, stall innovation, and mistake activity for productivity. Organizations that treat roles as deliberately designed systems, governed with the same seriousness as budgets, can capture the legitimate benefits of leaner structures and new technology without breaking the people who do the work. The choice is not between efficiency and wellbeing. It is between designing work on purpose and letting it design itself.
Subtract first: Before redistributing any work, inventory it and decide what to stop, simplify, or automate; make removal a recurring discipline with an owner.
Re-charter changed roles: After any restructuring, document the purpose, responsibilities, and decision rights of every significantly changed role, including peers who absorbed work informally.
Budget capacity: Require initiatives, including AI deployments, to state whose time they will consume and what will be deprioritized.
Count hidden work: Allocate time for AI experimentation and collaboration, and recognize it in performance and promotion systems.
Watch the distribution: Monitor who absorbs redistributed work, with attention to high performers, helpful colleagues, and women carrying low-promotability tasks.
Resource managers: Pair wider spans with reduced administrative load and real support, and treat manager exhaustion as a design signal.
Enable crafting within redesign: Give employees structured permission to reshape their roles, backed by systemic changes so that crafting does not simply shift burden.
Check the view from the top: Test leadership assumptions about efficiency gains against what front-line employees actually experience.
Research Infographic
References
Asana. (2021, January 14). Asana Anatomy of Work Index 2021: Work about work is dominating in a distributed world [Press release]. Business Wire.
Ashton, M. (2018). Getting rid of stupid stuff. New England Journal of Medicine, 379(19), 1789–1791.
Babcock, L., Recalde, M. P., Vesterlund, L., & Weingart, L. (2017). Gender differences in accepting and receiving requests for tasks with low promotability. American Economic Review, 107(3), 714–747.
Bakker, A. B., & Demerouti, E. (2007). The job demands-resources model: State of the art. Journal of Managerial Psychology, 22(3), 309–328.
Bakker, A. B., & Demerouti, E. (2017). Job demands-resources theory: Taking stock and looking forward. Journal of Occupational Health Psychology, 22(3), 273–285.
Bolino, M. C., & Turnley, W. H. (2005). The personal costs of citizenship behavior: The relationship between individual initiative and role overload, job stress, and work-family conflict. Journal of Applied Psychology, 90(4), 740–748.
Bove, T. (2023, January 3). Shopify is axing all meetings involving more than two people in a remote work "calendar purge" that the company itself calls "fast and chaotic." Fortune.
Bowling, N. A., Alarcon, G. M., Bragg, C. B., & Hartman, M. J. (2015). A meta-analytic examination of the potential correlates and consequences of workload. Work & Stress, 29(2), 95–113.
Colvin, C. (2026, September 10). Korn Ferry: Leaders have to rethink how work gets done. HR Dive.
Cross, R., Rebele, R., & Grant, A. (2016). Collaborative overload. Harvard Business Review, 94(1), 74–79.
Fischer, S. (2023, March 14). Meta to cut 10,000 more jobs, eliminate 5,000 openings. Axios.
Future Forum. (n.d.). Building a "Virtual First" culture from the ground up [Case study]. Future Forum.
Hackman, J. R., & Oldham, G. R. (1976). Motivation through the design of work: Test of a theory. Organizational Behavior and Human Performance, 16(2), 250–279.
Hogg, R. (2024, December 24). A year after Bayer got rid of bosses and asked staff to self-organize, its CEO says attrition has fallen. Fortune.
Kahn, R. L., Wolfe, D. M., Quinn, R. P., Snoek, J. D., & Rosenthal, R. A. (1964). Organizational stress: Studies in role conflict and ambiguity. Wiley.
Karunakaran, A., Kellogg, K. C., & Wiesenfeld, B. (2026, August 6). AI experiments need domain experts. Here's how to support them. Harvard Business Review.
Korn Ferry. (2026, September 8). Korn Ferry Workforce 2026 report: Unlocking growth requires rethinking how work gets done [Press release]. Korn Ferry.
Lake, S. (2024, November 8). Amazon CEO Andy Jassy hates bureaucracy so much he's planning to increase the ratio of employees to managers. Fortune.
LePine, J. A., Podsakoff, N. P., & LePine, M. A. (2005). A meta-analytic test of the challenge stressor-hindrance stressor framework: An explanation for inconsistent relationships among stressors and performance. Academy of Management Journal, 48(5), 764–775.
Lin, B. (2026, October 1). The rise of the "two-job job": 61% of workers say they're now doing more than one role. Fortune.
Maslach, C., & Leiter, M. P. (2016). Understanding the burnout experience: Recent research and its implications for psychiatry. World Psychiatry, 15(2), 103–111.
Microsoft. (2025, June 17). Breaking down the infinite workday [Work Trend Index special report]. Microsoft WorkLab.
Morgeson, F. P., & Humphrey, S. E. (2006). The Work Design Questionnaire (WDQ): Developing and validating a comprehensive measure for assessing job design and the nature of work. Journal of Applied Psychology, 91(6), 1321–1339.
Parker, S. K., Morgeson, F. P., & Johns, G. (2017). One hundred years of work design research: Looking back and looking forward. Journal of Applied Psychology, 102(3), 403–420.
Perlow, L. A. (1999). The time famine: Toward a sociology of work time. Administrative Science Quarterly, 44(1), 57–81.
Reuters. (2023, November 21). Citigroup CEO announces management changes, "difficult" decisions. Malay Mail.
Rizzo, J. R., House, R. J., & Lirtzman, S. I. (1970). Role conflict and ambiguity in complex organizations. Administrative Science Quarterly, 15(2), 150–163.
Sarre, S., Maben, J., Griffiths, P., Chable, R., & Robert, G. (2019). The 10-year impact of a ward-level quality improvement intervention in acute hospitals: A multiple methods study. Health Services and Delivery Research, 7(28).
Schaal, D. (2026, September 2). Uber to lay off 3,300 employees, reducing management jobs by 20%. Skift.
Star, S. L., & Strauss, A. (1999). Layers of silence, arenas of voice: The ecology of visible and invisible work. Computer Supported Cooperative Work, 8(1–2), 9–30.
Tims, M., Bakker, A. B., & Derks, D. (2012). Development and validation of the job crafting scale. Journal of Vocational Behavior, 80(1), 173–186.
Wrzesniewski, A., & Dutton, J. E. (2001). Crafting a job: Revisiting employees as active crafters of their work. Academy of Management Review, 26(2), 179–201.
Jonathan H. Westover, PhD, Chief Research Officer (Nexus Institute for Work and AI); Co-Founder & Chief Workforce and Learning Officer (Future State University); Founder & CEO (Human Capital Innovations); Professor of Organizational Leadership & Change (UVU). Read Jonathan Westover's executive profile here.
Suggested Citation: Westover, J. H. (2026). The Two-Job Job: Role Sprawl, Hidden Workload, and How Leaders Can Redesign Work Before It Breaks People. Human Capital Leadership Review, 40(3). doi.org/10.70175/hclreview.2020.40.3.3
