The return-to-office debate is more than a disagreement about commuting or employee preferences. It exposes a deeper management question: does the way an organization engages knowledge workers still fit the way its work is produced?
Information and communications technology has loosened the connection between productive contribution, a particular location, and a shared working schedule. AI adds another uncertainty: which contributions will continue to require people, and how many?
For management, this is not simply a workplace-policy problem. It is an operating-model problem, with implications for employment, organizational capability, offices, and business travel.
The inherited model bundles together things that can now be separated
The conventional office-based model brings several arrangements together. The organization employs people, assigns responsibilities, establishes working hours, and requires attendance at a designated workplace.
That bundle provides a familiar structure for supervision and coordination. But its components need not remain inseparable.
A worker can be available without being in the office. A team can interact in real time without sharing a location. People can contribute at different times to a common undertaking. Physical gathering can serve a particular purpose without becoming a daily requirement.
The strategic question is therefore not simply, “How many days should employees attend?”
It is, “Which parts of our work require shared time, shared place, continuing employment, or a defined project arrangement?”
Start with the activity, not the attendance policy
An attendance-first approach establishes where people must be, then organizes activities within that requirement.
An activity-first approach examines the work and establishes the conditions it needs.
This does not mean eliminating offices or assuming that every task should be remote. It means distinguishing among concentrated individual work, routine coordination, exploratory collaboration, learning, relationship development, and client engagement.
Research illustrates why that distinction matters. Experiments found that videoconferencing inhibited creative idea generation, but did not make groups less effective at selecting which ideas to pursue.1 “Collaboration” is therefore too broad a category to justify a single location rule.
A useful test for gathering is:
What becomes possible or materially better because these particular people are together, and is that benefit worth the travel, time, and disruption?
Brainstorming, apprenticeship, resolving disagreements, and social reinforcement can all be legitimate purposes. They should be designed as such, rather than left to justify undifferentiated attendance after the fact.
RTO tension reveals a missing management agreement
Management has legitimate responsibilities for performance, coordination, organizational learning, and accountability. Staff have legitimate interests in avoiding unnecessary travel, protecting concentrated working time, and understanding why attendance is required.
The conflict becomes difficult to resolve when the two sides are discussing different things.
Staff may be asking whether their activities require physical presence. Management may be asserting its authority to establish attendance as a general condition of employment.
Neither an employee’s individual productivity figures nor a general appeal to “culture” necessarily resolves that disagreement. What is missing is an explicit agreement about how the organization will coordinate work, develop capability, and assess contribution.
Asynchronous work makes that agreement particularly important. It needs clear decision ownership, dependable handoffs, accessible records, response expectations, and ways to identify blocked work.
The alternative to attendance-based coordination is not the absence of structure. It is a different structure, designed deliberately.
Employment itself belongs in the redesign
The deeper issue is not whether employees can work from home. It is which continuing commitments make employment the appropriate relationship in the first place.
A useful redefinition for knowledge work is:
Employment is a continuing relationship through which an organization secures expertise and capacity for evolving responsibilities, while providing compensation, resources, security, and agreed limits on that commitment.
Under this definition, continuity and adaptability justify employment. Physical attendance does not define it.
Project contracting serves a different purpose: a person or group accepts responsibility for a bounded undertaking or agreed result. The group may organize its own methods, schedule, location, and contributions.
Organizations could consequently use a combination of continuing employees, persistent specialist groups, and autonomous project teams. The appropriate mix would follow from the responsibilities involved, rather than a presumption that every capability must be staffed internally.
This also changes the meaning of working time. Defined hours can establish the boundary of an employer’s claim and the availability needed for coordination, rather than serve as proof that productive work has occurred.
Removing attendance requirements must not create an unlimited obligation to be available. Nor should outcome-based assessment overlook experimentation, mentoring, or maintaining capabilities that matter beyond the next deliverable.
AI makes the staffing question unavoidable
AI introduces a second dimension of change: not only where knowledge work happens, but how much human participation particular activities require.
There are warning signs worth monitoring. Stanford’s August 2026 employment analysis found that young workers in highly AI-exposed occupations were falling behind less-exposed peers, primarily through reduced hiring rather than increased separations. The researchers did not find widespread economy-wide displacement and described their findings as indicators rather than causal estimates.2
For management, the implication is to plan across several possibilities: augmentation of existing teams, lower staffing requirements, new service demand, and changing pathways into professional work.
A smaller employed workforce does not automatically create a thriving project market. Automation could reduce demand for work previously purchased from contractors as well.
The planning challenge is to preserve the expertise, learning pathways, and accountability an organization needs while changing how work is produced.
A practical agenda for management
Over the coming decades, no single attendance policy or staffing forecast will provide a durable answer. A structured advisory process can help management connect decisions that are too often made separately:
- Map the work. Identify human contributions, dependencies, timing requirements, and the purposes of physical gathering.
- Design coordination. Establish decision rights, documentation, handoffs, availability, and accountability.
- Choose engagement arrangements. Determine which capabilities require continuing employment and which suit external groups or projects.
- Align spaces and travel. Make office requirements and gathering budgets consequences of the work model.
- Test and adapt. Evaluate arrangements against performance, learning, client outcomes, cost, and worker experience.
Geography follows these choices. Weekly attendance keeps workers tied to a practical commuting region; occasional gatherings permit a wider residential and recruitment range. Location policy is therefore also a talent and organizational-design decision.
The next management challenge is not restoring the office or abolishing employment. It is establishing what each is for.
Organizations need help making sense of these changes before committing to policies, staffing structures, and property decisions that may reinforce yesterday’s assumptions. The starting point is a better question: what work must we accomplish, what continuing responsibilities must we retain, and what arrangements best support both?
References
- Brucks, Melanie S., and Levav, Jonathan (2022). Virtual communication curbs creative idea generation. Nature, 605(7908), 108-112. https://doi.org/10.1038/s41586-022-04643-y
- Stanford Digital Economy Lab employment research summary