Quantitative data forms the basis for decisions in the multi-billion-dollar commercial real estate sector. Ultimately, however, a multitude of qualitative data determines the success of an investment. The answers to a successful investment were never limited to tangible, quantitative market or performance data such as rents or occupancy rates, appraised property values, comparable sales figures, or the projected return on investment or internal rate of return of an asset.
Qualitative assessments may be elusive, yet they determine how a property is perceived and seal the fate of an investment. Such data offers insight into the past and present history of a property's environment and is invaluable for understanding how an investment's lifecycle might unfold. The stories told by this information can shape risk profiles and ultimately dictate where capital is deployed and which markets are too quickly written off.
Artificial intelligence is changing how the commercial real estate industry can collect this information, and market participants would do well not to overlook it. In New York, a developer can change a building's name or address to gain a slight advantage. For instance, when owners renamed 708 Third Ave to 10 Grand Central or 1250 Broadway to NoMad Tower, they altered the perception of these properties in the market and achieved significantly higher rents.
Many smaller investors and developers across the country often do not have the luxury of simply renaming a building. In most cases, an investment depends on a variety of qualitative factors that outline the underlying history of a property, neighbourhood, or submarket. Historically, finding and analysing such information was arduous, if not inaccessible, and relied on intuition, which increased underwriting risk. It also wasn't scalable. New technologies are changing this.
Much of the data needed to reliably identify qualitative risks in real estate investments is clearly present. It can be found in historical news reports, environmental assessments, permit applications, local ordinances, public transport plans, crime data, or the minutes of zoning committee hearings, school board meetings, or neighbourhood forums. These factors can, among other things, provide measures of current and future economic activity, neighbourhood safety, public transport accessibility, or even school quality.
While this information was traditionally dismissed as 'soft sentiment' because it couldn't be easily transferred into a spreadsheet, advanced analytical tools can now process these disparate records. It is finally becoming possible to convert elusive neighbourhood signals into hard data that underwriters can actually test.
The consequences of failing to find and test these stories are particularly evident in markets where narratives evolve faster than fundamentals. Austin and San Francisco offer two very different examples in recent years of what happens when investors allow a narrative to influence underwriting too strongly. In Austin, the pandemic created a strong narrative around the city's future growth, as corporate relocations, tech expansion, population growth, and investor enthusiasm reinforced each other. However, belief in this growth justified a wave of speculative office development that the market could not absorb. Developers built nearly 14 million square feet of new office space in Austin between 2020 and the end of 2025, pushing the city-wide vacancy rate to 29 per cent and the downtown vacancy rate to 32.4 per cent.
San Francisco shows the other side of the same problem. Around 2022, the 'urban doom loop' narrative became the dominant story about the city. Office vacancies rose, retailers left the city centre, and public safety concerns made national headlines. By 2025, the market showed that relying on this narrative meant missing an opportunity. San Francisco recorded its strongest year for office leasing since 2019, ending last year with over 1 million square feet of positive net absorption in the fourth quarter. Although the city still faces challenges, the narrative lagged behind the signals that were actually generating value. In Austin, an oversupply chased a growth story, and in San Francisco, real estate investors risked missing a recovery because the decline story seemed permanent. Both examples demonstrate how narratives can lead investors to short-sighted decisions.
In summary, real estate investors tend to value based on 'vibe', sometimes overlooking the value of qualitative data that can be the 'secret ingredient' for investment success. Yet, in a data-rich, multi-billion-dollar industry like commercial real estate, should so much judgment still rely on instinct? Used cautiously, AI can help transform softer, more elusive signals into something that underwriters can test. It can scan public meetings for recurring concerns about safety or infrastructure, compare how often local businesses open or close, track whether permitting activity is concentrated around a corridor, or identify when tenants and brokers describe a neighbourhood differently than six months prior. This information gives underwriters a way to see if the story around a neighbourhood aligns with the underlying data. San Francisco's office market shows why this is important. The story investors often tell about office demand is relatively simple: the newest buildings with the highest amenities in the most established business districts will win. However, some of today's fastest-growing tenants complicate this assumption. In San Francisco, AI start-ups are increasingly concentrating in smaller, mixed-use neighbourhoods and converted industrial buildings, rather than defaulting to established office towers.
