By: General Manager Caroline Stoll, Data & Analytics at LightBox
Artificial intelligence (AI) is rapidly infiltrating the world of Geographic Information Systems (GIS). From automating map analysis to powering predictive models, AI-driven technologies are touted as the next revolution in geospatial technology. But is the reality living up to the hype? Many current implementations fall short of their lofty promises.
In commercial real estate (CRE), where location intelligence is the foundation of investment, development, and risk strategy, these blind spots carry significant implications. This article challenges the GIS and commercial real estate (CRE) industries to demand more from GeoAI—more context, more transparency, more relevance, and more responsibility.
While early applications of GeoAI show promise, many still fall short—held back by gaps in relevance, usability, and trust. To move beyond the hype, the industry must confront the biggest barriers standing in the way.
Here are four major blind spots—and how we can begin to address them.
1. Stuck in the Present: The Need for Temporal GIS Data
In CRE, location isn’t just about where—it’s about when. Trends like zoning changes, infrastructure expansion, and land-use evolution play out over years. Yet most AI models treat CRE data as static, ignoring the temporal dynamics critical to forecasting property value or identifying risk exposure. They analyze ‘snapshots’ of the world without incorporating the dynamics of time. A survey of geospatial data mining research noted: “The challenge stems from the difficulty of temporal representation on maps because of the limitation of GIS in representing dynamic processes.” This blind spot limits predictive power and strategic foresight.
Consider climate resilience or long-range investment planning. If an AI model ignores accelerating development or environmental change, it misjudges future opportunity and risk. As the World Economic Forum puts it: “The past is no longer a reliable predictor of the future.”
Forward-thinking platforms—such as those that incorporate historical permit data, ownership turnover, and aerial imagery—are stepping into this gap. LightBox, for example, is investing in deep, time-aware property graphs to surface meaningful temporal patterns. The goal: illuminate not just what’s happening today, but what’s likely to happen tomorrow.
2. The Ownership Black Box: Why Land Use and Transparency Matter
Location intelligence is only as good as the data behind it. In CRE, few data points matter more than land use and ownership—who controls that parcel, that building, that asset, and what it’s intended for. While GIS and geospatial AI are powerful across many domains, ownership opacity significantly limits their value in CRE, where understanding both the “who” and the “what” behind property holdings is essential to investment decisions, risk analysis, and policy enforcement.
Property records might list an LLC or trust as the owner, with the real beneficiary obscured behind layers of shell companies. This lack of transparency isn’t just a legal or ethical headache—it creates blind spots in AI models built to understand land aggregation, redevelopment trends, or investor behavior. Practices like shell-company land banking can distort markets and sideline communities.
The Brookings Institution has highlighted the scale of this problem: “Offshore investors can easily hide their identity by using opaque corporate ownership structures… [and] because this practice is so common, offshore investment in the real estate sector is likely far greater than what can be measured with public data.” Fortunately, regulatory reform is catching up. In the U.S., new laws are requiring LLCs to disclose beneficial ownership. In parallel, platforms like LightBox are building AI-powered ownership graphs to reveal hidden relationships between parcels, investors, and development patterns. The goal: help CRE professionals see through the maze and make smarter, faster decisions.
3. Between Hype and Reality: Aligning AI with Real-World GIS Workflows
Few technologies have been as hyped in recent years as artificial intelligence. In GIS conferences and marketing materials, we hear how AI will automate mapping, drive smart cities, and solve problems from traffic to deforestation. However, on the ground, adoption of AI in geospatial workflows remains limited. GIS professionals often find that flashy AI algorithms aren’t addressing their most pressing needs, or that the solutions are too complex to integrate into existing systems and processes.
Only 2% of local governments around the world reported using AI in any capacity—despite widespread interest, according to a Bloomberg Philanthropies 2023 global survey. Many AI initiatives fail to move beyond pilot stage or show clear ROI. In geospatial sectors, departments still rely on basic GIS analysis and manual workflows.
To overcome the disconnect, the next generation of GeoAI must be grounded in real-world applications and designed for the people who use them. This means co-developing AI tools with planners, engineers, and scientists. LightBox, for example, focuses on integrating AI insights into familiar GIS interfaces and aligning AI with everyday GIS tasks like automating data cleaning or identifying changes in property conditions.
4. Data Integrity and Ethics: AI in Geospatial Analytics
Even the most sophisticated AI is only as good as the data it’s fed and the context in which it’s deployed. In GIS, this raises concerns about data integrity (accuracy, completeness, reliability) and ethical use (fairness, transparency, and respect for privacy). Both expose major blind spots.
Take broadband coverage maps in the U.S. for example. Microsoft found that the FCC’s map vastly undercounted the number of people without access to high-speed internet, showing 25 million underserved versus their own estimate of over 162 million. Bad data led to misdirected funding and millions being left behind. This is a prime example of a data integrity failure with ethical implications.
On the ethics side, AI in GIS can inadvertently perpetuate biases. If training data is skewed or black-box models can’t explain their decisions, the risk of inequitable outcomes increases. Companies like LightBox are addressing this by investing in data stewardship, transparency, and user controls. These measures ensure that GIS professionals remain in the loop, with clear, explainable outputs that reflect policy values and practical realities.
The Way Forward: Demand More from GeoAI
Yes, we’ve seen promising advances in AI for GIS. But significant blind spots remain: lack of temporal context, ownership opacity, disconnect from real-world workflows, and questions of integrity and ethics. These aren’t failures—they’re opportunities to improve.
In CRE especially, decisions hinge on nuance—location, history, ownership, infrastructure, and risk. Professionals must demand AI that provides context, not just output. Insist on transparent data, credible lineage, and meaningful integration. Ask vendors: how will this tool make my planning, underwriting, or land use analysis better? If they can’t answer, move on.
The encouraging news is that companies like LightBox are leaning into these challenges. By investing in time-aware data structures, transparent ownership mapping, and integrity-first design, they’re creating GeoAI solutions that work where it counts. Let’s build on that momentum. The next generation of AI in GIS—and in CRE—shouldn’t just be exciting. It should be accountable, insightful, and built to serve those making the most consequential decisions about land, investment, and infrastructure.
This blog is part of a special series leading up to the Esri User Conference, where we’re sharing key insights and trends shaping the GIS industry.
Are you attending the Esri User Conference this year? We’d love to connect and hear what topics you’re most excited about! Drop us a note at hellouc@lightboxre.com—let’s chat.
Subscribe To Insights
Be the first to get the latest news, updates and insights from LightBox.
