Can AI Help Predict Office Space Demand After Remote Work Shifts?

Measuring What’s Happening, Which Is Different From Predicting What’s Next

AI has become a genuinely useful tool for understanding how office space is actually being used right now — badge access records, occupancy sensors, and desk booking data can all be analyzed to show real patterns in attendance, like which days of the week see the most people in the office or how utilization varies by department or floor. This kind of analysis gives companies and commercial landlords a much clearer, evidence-based picture of current space usage than relying on assumptions or infrequent manual observation.

Where this gets harder is moving from measuring current utilization to reliably forecasting future demand. Office space demand isn’t purely a function of gradual market trends the way some other real estate categories are — it’s heavily shaped by discrete corporate policy decisions about remote and hybrid work, and those decisions can shift relatively suddenly and vary enormously from one company or industry to another.

Why Company-Level Policy Decisions Complicate Forecasting

A retail site’s foot traffic or a residential market’s price trend tends to move somewhat gradually and follows patterns that historical data can meaningfully inform. Office demand is different because a single large employer’s decision to shift its work-from-home policy can change its own real estate needs substantially and relatively quickly, and that decision is driven by internal corporate factors — company culture, industry norms, leadership preferences — that don’t necessarily follow a predictable pattern an AI model trained on historical real estate data would capture well.

This is a genuine structural challenge for AI forecasting in this specific corner of commercial real estate: the input that matters most is often a policy choice happening inside individual companies, not a market-wide trend visible in aggregated real estate data alone.

How AI Data Still Adds Real Value Despite the Forecasting Challenge

Even without solving the harder problem of precisely predicting future demand, AI-generated utilization data has real practical value for near-term decisions. Companies can use it to right-size their existing footprint based on actual, current usage rather than guesswork, and landlords can use aggregated utilization trends across their portfolio to inform decisions about reconfiguring space or investing in amenities that appear to actually drive in-person attendance. This is a more modest, grounded application than trying to forecast market-wide office demand years into the future, but it’s one where the technology delivers clear, measurable value today.

Bottom Line

AI is genuinely useful for measuring how office space is currently being used, giving companies and landlords real data to inform near-term decisions. Predicting future office space demand after remote work shifts remains a much harder problem, because it depends heavily on individual company policy decisions that don’t follow the kind of consistent historical pattern AI forecasting tools rely on.

Go deeper

• Can AI Optimize Product Assortment for Individual Store Locations?
• What Data Sources Feed AI Demand Forecasting Models?