The Return-to-Office Debate Has a Customer-Shaped Hole in It

Nobody asked the customer. That's the thing. We have spent five years locked in this furious, occasionally hysterical debate about whether people can be trusted to work from home — and at no point did anyone think to ask the one question that actually decides the whole thing.

Not "Can workers be productive from home?"

But: "Do customers still want to come to your office?"

Because the answer to that question is no. Emphatic, permanent, load-bearing no. And once you accept it — once you really let it settle — roughly half of the working-from-home argument collapses. You're not debating policy anymore. You're describing something that already happened while everyone was arguing about whether it should.

The Customer Already Left. You Weren't Consulted.

Think about what an office was for.

Not the ping-pong tables. Not the motivational posters. An office existed, first and foremost, as a place where customers came to interact with you. A bank branch. An insurance office. A government department. A law firm. The office was the interface — the physical membrane between the organisation and the people it served.

Now think about when you last voluntarily walked into one of these places in a good mood.

You didn't. Nobody does.

The default position of the modern customer is: I will do everything online, through an app, via a chatbot, or over the phone — and I will only physically visit an office as an absolute last resort, when something has gone horribly wrong, at which point I will arrive furious.

This is a massive shift and we barely talk about it. The customer's relationship with the physical office has moved from "routine" to "emergency-only." The branch visit has become the customer equivalent of the emergency room — you only go when things are truly broken, and you're not in the mood for small talk when you get there.

So when a CEO bangs on about "getting everyone back in the office," I want to ask a simple question: Back in the office to do what, exactly? To sit in a building that the customers have already abandoned?

You're not preserving a way of working. You're maintaining a theatrical set.

Bank of America has closed over 800 branches since 2017. JPMorgan Chase, hundreds more. The IRS reports that over 90% of tax returns are now filed electronically. Insurance claims, mortgage applications, customer support — all migrated online, not because companies were feeling generous, but because customers demanded it by simply refusing to show up.

Roughly half of what we call "working from home" wasn't a decision made by workers at all. It was dictated by customers. They voted with their browsers. They chose the app over the appointment, the chatbot over the queue. The debate is over. They're not coming back. You might as well argue about whether people will return to renting DVDs from Blockbuster.

Now Let's Talk About the Other Half

The remaining portion — the part that genuinely is about whether employees can perform complex, collaborative work remotely — deserves a more honest answer than it usually gets.

The common objection goes like this: "You can't do serious, coordinated work from home. You need to be in the same room."

To which I'd like to introduce you to the United States military.

The U.S. military now conducts a remarkable proportion of its operations remotely. Drone operators in Nevada execute missions over Yemen. Intelligence analysts in Maryland coordinate with special forces across three continents. The most complex, high-stakes, zero-margin-for-error operations in human history are run by people who are not in the same building, the same city, or even the same hemisphere as the action.

Now. Please explain to me what it is about preparing a quarterly brand review or coordinating a product launch that is so extraordinarily complex that it requires physical co-location, when the task of conducting military operations across fourteen time zones apparently does not.

The answer is nothing.

The military can do it because they were forced to build the systems, the protocols, and the discipline to make remote coordination work. Corporations haven't been forced yet, so they haven't bothered. That's not an argument against remote work. That's an indictment of corporate laziness.

The Real Problem: The Software Is Still Prehistoric

Here's where the actually interesting problem lives. The part that nobody's talking about.

The reason remote work often feels unsatisfying isn't because human beings need to be in the same room. It's because the software we use for remote collaboration is at the cave painting stage of its evolution.

Consider Zoom.

What is Zoom, really? It's a blank canvas. A rectangular window through which you can see someone's face and hear their voice. That's it. No templates. No task structures. No workflows. No embedded decision-making frameworks. Nothing that tells the system what kind of conversation you're having or what should happen next.

Now think about how far text-based business software has come.

We have databases, CRM systems, project management platforms, form builders, workflow engines, ERP systems — decades of evolution in structured, text-based tools that encode business logic, enforce processes, and guide human behaviour. Salesforce alone has thousands of features designed to structure how humans interact around business tasks.

Now look at video and audio communication software.

It's a phone call with a camera. We took a technology that Alexander Graham Bell would recognise and added a webcam. That's it. That's what we did. We patted ourselves on the back and moved on.

Let me put this in terms any developer will understand instantly.

Imagine if the entire history of text-based business software had stopped at Notepad.

Not Word. Not Excel. Not Salesforce or SAP or Jira. Just a blank screen where you could type characters. No templates, no fields, no validation, no formulas, no workflows. Just a cursor blinking on an empty page.

Need to do accounting? Open Notepad. Need to manage a thousand customer relationships? Notepad. Need to coordinate a supply chain across six countries? Here's a blank text file. Good luck.

You'd look at that and say: "This is obviously inadequate. Text as a medium isn't the problem — the tools are the problem. We haven't built the structures yet."

And you'd be right. That's exactly what happened. The text world went from blank files to databases, then to forms, then to spreadsheets, then to CRM systems, then to ERP platforms, then to workflow engines with task templates and automated triggers and role-based access controls. Forty years of relentless evolution from "blank page" to structured, business-logic-aware environment.

Zoom is Notepad.

That's the whole diagnosis.

It has no concept of what kind of meeting you're in. It doesn't know if this is a sales call or a board review or a customer complaint or a weekly standup. It has no tasks, no templates, no embedded decision frameworks, no structured follow-ups. It's the equivalent of running your entire enterprise on a blank text file and then concluding that screens are inferior to paper.

We would never look at Notepad and say, "Text-based work has clearly failed as a concept." We'd say, "We need better tools."

Yet that is precisely the logical error we make with remote video work every single day. We look at Zoom — our Notepad, our blinking cursor on an empty screen — and conclude that the medium is the problem.

It isn't. The tooling is the problem. We just haven't built the structures yet.

The Only Part of Capitalism That Actually Matters

"Capitalism is not a process of efficiency optimization; it is a process of discovery."

Before I get to the technology that could fix this, I need to say something about what we're doing with technology in general right now. Because I think we're making a catastrophic error, and it's not a technical one. It's a philosophical one.

Capitalism has exactly one genuine justification for its existence: discovery. Not efficiency. Not cost reduction. Not quarterly earnings. The creation of things that didn't exist before. The invention of new value. That's the engine. That's the part that makes the whole messy, unequal, sometimes brutal system worth defending — because when it works, it produces things that genuinely make human life better in ways no committee would have planned.

Everything else capitalism does — the cost-cutting, the optimising, the restructuring, the endless trimming of headcount — is just accounting. Important accounting, sure. But accounting nonetheless. You can't cost-cut your way to a future anyone wants to live in.

And right now, at the exact moment when a genuinely transformative technology has arrived — one that could create entirely new categories of human collaboration — the corporate world has looked at it and seen a cost-cutting tool. We are pointing AI at the accounting part of capitalism and completely ignoring the innovation part. Faster Jira. Cheaper Salesforce. An ERP system that needs fewer developers to maintain. We have a technology that could reinvent how human beings communicate and coordinate across distances — the single biggest unsolved problem in modern work — and we're using it to shave 15% off the cost of tools we built in the 1990s.

That's not strategy. That's a failure of imagination so complete it's almost impressive.

Because the gap I just described — between Notepad-era video software and the structured, intelligent communication tools we actually need — is precisely the kind of gap AI was born to close. Not in forty years. In about five.

Think about what AI can do inside a video call. It can understand context — recognise whether you're in a sales meeting, a support call, or a project review. It can generate real-time agendas, capture action items, surface relevant documents, flag unresolved decisions, prompt follow-ups. All without anyone typing a word.

It can turn a Zoom call from Notepad into something closer to Salesforce — automatically — by understanding what's happening and building the structure on the fly.

But it goes further.

AI doesn't just take notes. It can understand the meeting better than the participants do. It can notice that three people said contradictory things about the timeline and nobody caught it. It can identify that a decision was implicitly made in minute fourteen but never explicitly confirmed. It can detect that the customer's tone shifted from relaxed to concerned at the twenty-two-minute mark and flag it for follow-up.

This isn't "text-based software but cheaper." This is an entirely new category. The kind of structured, context-aware environment that video communication has needed since the first Zoom call stuttered into existence.

The corporations busy using AI to trim costs on their existing systems are about to discover that AI's real value isn't making the old world cheaper. It's making the new world possible.

The companies that understand this will build remote-work tools so good that the "return to office" debate will sound as dated as arguing about the merits of the fax machine. The companies that don't will have a slightly cheaper CRM and a half-empty building they're still paying rent on.

I know which architecture I'd bet on.