Algorithms Don’t Need an Office
Those who make their living in the office market, investors, lenders, and brokers, tend to be averse to change. Big shifts, such as the pandemic, trigger a collective head-in-the-sand moment. Here in San Francisco, despite being in the epicenter of AI, it’s difficult to find thoughtful real estate industry dialogue about how AI might reduce the labor force and lower demand for office. It’s a bit like contemplating one’s own demise.
To be sure, the surging demand for office space from AI companies makes for an awkward bedfellow. On one hand, market players welcome an expansionary sector — it’s exactly what we need to fill our empty buildings. On the other, we quietly worry demand for office may surge while AI is being built but rapidly decline as AI is deployed and begins replacing workers at scale.
For my part, I’ve tried to imagine how commercial office brokerage might change with AI. It’s easy to see the obvious stuff but difficult to look ahead 10 years. Recently, I asked a young professional who straddles tech and real estate what a firm like Cushman & Wakefield might look like in ten years. Without hesitation, he said, “It’ll be a branded AI agent,” which sounded to me like an app. He envisions large CRE firms evolving into AI-powered agents based on proprietary data and analytics. What’s more, these agents will interact (largely) with other AI agents. A set of desired outcomes will be provided to an agent which will navigate other agents to fulfill the objectives. Less humans. Is this a reasonable view of the future? I have no idea. But here’s what I do know: algorithms don’t need an office.
Given my challenges visualizing the future of my industry, I decided to tackle a less complex subject: food. My family’s food consumption has gone decidedly digital. Date night? OpenTable. Groceries? Instacart. Dinner at home? Often Doordash. These platforms are guided by AI which tracks our habits, prompting reorders, suggesting meals before I even think about them.
For now, people are still in the mix. My Instacart orders are fulfilled by human shoppers, my Doordash meals delivered by drivers. But how long will that last? Imagine Amazon introduces a smart refrigerator. Call it the hardware (like my Peloton bike). For an annual subscription, Amazon delivers me the fridge at no cost. It monitors our food levels, links to Amazon’s robotic warehouses (and my bank account), and automatically replenishes my inventory. It suggests new recipes and with a click the recipe and ingredients are at my fingertips. Delivery is fulfilled by a robot at the warehouse loading an autonomous vehicle or drone. No store. No people. Just AI, logistics, and robotics.
Once Amazon controls product acquisition and delivery, why not also control its production? Amazon could take over farming, employing advanced robotics and bio-engineered solutions. At this point, they control the entire vertical, from inception to delivery.
Some AI evangelists argue we’re heading to a great place in which we’ll be freed from many burdensome tasks. However, in many cases, these tasks are also known as “work.” Work for which people receive compensation. So, one important question: what do we do when the performance of our business no longer requires us?
Optimists also liken this moment to the Industrial Revolution, which shifted the economy from agrarian to manufacturing and ultimately raised living standards. Machines replaced some jobs but created others. Perhaps AI will do the same. But if new forms of work emerge, will it happen in offices?
If I were a major office investor or developer, I’d be talking regularly with AI leaders, computer scientists, and futurists to learn as much as possible. But I’m just one guy trying to see around the corner of my own small world. For now, I’m keeping an eye out for that Amazon refrigerator.