Companies are dragging everyone back into the office to preserve the “intangibles” of human interaction. At the same time, they are rolling out AI tools whose entire purpose is to reduce the need for human interaction. Seems like a contradiction to me.
During Covid, most tech work moved from the office to the home. It turned out that almost everyone was happier. No fighting traffic, parking or public transit. No crowded meetings, kitchens or bathrooms. It turns out you can type on a keyboard at home just as easily as you can in a cubicle.
For the smaller group of people who still needed to go in (me included), things improved too. Offices were quieter, traffic was lighter and shared resources were available. From what I saw, productivity improved across the board.
The RTO-AI Contradiction
Companies are mandating return-to-office (RTO) policies to preserve “intangible” human interactions while simultaneously deploying AI tools that reduce the need for those very conversations.
-
The RTO Argument: Physical proximity fosters serendipitous learning, mentorship, and deep knowledge sharing between junior and senior staff.
-
The AI Impact: AI-driven workflows automate the starting point of tasks — the exact stage where human collaboration and architectural judgment are most critical.
-
The Result: Interaction becomes shallower and more tactical, eroding the internal relationships and expertise that RTO policies claim to protect.
What Is the Purpose of RTO?
Despite this (and despite employee preferences) many companies decided they wanted workers back in the office. To justify it, they leaned on the idea of “intangible benefits.” The argument was that physical proximity creates serendipitous interactions that generate value for the company.
“Intangible benefits” is a convenient position because it isn’t falsifiable. Productivity can be measured, however. KPIs exist for a reason. But since no one can measure intangible benefits, the argument sidesteps evidence entirely.
That said, there is a kernel of truth here. Employee interaction matters. When someone completes a task, they have the deepest possible understanding of it. That knowledge includes not just how the process works, but why it works, why other approaches didn’t work and the entire path from start to end that was required to complete the project. That full, holistic understanding is extremely valuable. Sharing that knowledge among team members helps less experienced employees grow and allows teams to connect insights across domains in ways that can lead to real innovation.
The intangible benefits were both unmeasurable and contained just enough truth that counterarguments could be dismissed. The unspoken corporate position was essentially: There is an immeasurable benefit to in-person interaction, and we believe that benefit outweighs the measurable productivity gains of remote work. Did those corporate leaders believe in those benefits or was the argument just an excuse to force workers back in the office?
The Hypocrisy of Pairing RTO With AI
Now comes the hypocrisy.
Those same companies are now aggressively pushing AI workflows onto their employees.
AI automates mediocrity. It does not produce complete, polished outputs. Its work is typically used as a starting point, but that starting point is exactly where the most employee interaction is required. The start of a task is the point where an employee has the least knowledge and the most to gain from interaction with colleagues.
Before AI workflows, employees had no choice but to talk to one another. If junior engineers needed to modify an internal codebase, they had to collaborate with senior engineers. Those senior engineers would explain not just how the system worked, but why it was built that way. The junior employees learned the system and absorbed architectural judgment along the way.
Now a junior employee asks an AI tool to modify the code. And it usually doesn’t work. Eventually, the junior employee still must ask the senior engineer for help, but now that interaction is narrower. Instead of a full walkthrough of the system and its design decisions, they get a tactical fix to a specific problem. The interaction is shorter, shallower and far less educational.
AI Limits Collaboration Where It Matters Most
This isn’t unique to software either.
A marketing consultant tasked with learning about a new industry used to lean heavily on colleagues with prior experience. Those conversations built expertise and relationships. Now the consultant asks AI for an industry overview and simply sanity-checks the output afterward.
In both cases, employees gain less knowledge and build fewer internal relationships. The casual conversations that once sparked new ideas or unexpected opportunities never happen. The “intangible benefits” companies claimed were critical to justify return-to-office policies are quietly eroded by the very AI workflows those same companies are championing.
You can’t claim that human interaction is essential, mandate return-to-office policies to protect it and then simultaneously strip away the very moments where knowledge transfer and collaboration actually happen. AI isn’t the problem, but pretending it strengthens culture while it quietly shrinks real human connection is. If companies genuinely believe in the value of in-person work, they must be just as intentional about preserving deep learning, mentorship and shared understanding as they are about automating output.
Save In-Person Work for Value Drivers
We also need to acknowledge that not everything a company does must be perfect. Many things can be good enough or even just OK. If a company’s internal website for recording employee hours looks like a GeoCities site from the 90s, that isn’t really a big problem. If the website they present to clients looks like that same GeoCities site, that is a major problem. Companies need to prioritize and must choose areas where excellence is most critical. They need to decide which areas differentiate them from competitors and which actually maximize their value. They must focus on and invest in those areas.
A good balance doesn’t need to be complicated. In-person work should be reserved for the parts of the business where people learn vital judgment and skills, not just execute them. That means the part of the business that drives value does so by being better than your competitors. By definition, being better means a company can’t be average. Therefore, developing human knowledge is critical for doing better work than whatever an AI can churn out. That’s where you need in-person collaboration most.
Remote work and AI tools make sense for execution-heavy tasks where speed and consistency matter more than learning. If leaders can’t articulate which parts of their organization fall into each category, then RTO isn’t a strategy. Instead, it’s just a mandate.
The flip side of that coin is acknowledging that not all employees must feel invested to be productive. Some people want a career, while others just need a salary. Both are entirely valid approaches to work. For the employees who wish to get the work done in return for a paycheck, let them use AI and work from home. Make sure they deal with the non-critical areas of the business.
Treat Your Employees Like Adults
Simply said, treat employees as adults. For the employees working in and on critical parts of the business, the company promises a mutual investment, working to develop their skills and experience in return for the in-person work. If there aren’t clear benefits for the employee and the organization, then RTO was never about culture at all.
