Real Estate Is Insecure About Its Relationship With Data and AI

Artificial intelligence can be a godsend in collecting and analyzing data at commercial real estate firms. It can also jeopardize the security of that very data. 

Solving that conundrum is becoming increasingly important to both the firms and the vendors who want to service them. 

“Real estate is becoming as much a data business as it is a property business,” said Terry Keller, chief technology officer at Solon, Ohio-based global giant MRI Software. “I think that shift is forcing our industry in particular to rethink cybersecurity. AI is allowing real estate companies, whether you’re a vendor like me or a property owner, to monitor these massive volumes of tenant transaction building data in real time.”

That’s a positive, Keller said. AI can detect anomalies in the data that humans might miss. 

“AI allows us to look for suspicious access patterns or potential fraud inside of transactions,” he said. “So as we start to think about the industry as a data business, AI as that helper becomes that shield. But the point of the story here is that it also becomes a brand-new attack surface for real estate data.”

A plethora of proptech and general tech companies are focused on dealing with real estate data and security attacks.

Of the general companies, CrowdStrike and Microsoft are two of the most prominent that focus on broad enterprise cybersecurity and include AI-specific components. 

Other firms heavily in the real estate data security mix include Reco, an AI-focused SaaS security platform; Claroty, a cyber-physical system company for smart buildings and legacy business management systems; and vendors like MRI Software, which are building AI platforms designed around cybersecurity principles and have published AI governance controls. 

ButterflyMX, a video and biometric digital access company, and Inviscid AI, a physics-informed AI building intelligence platform for automated mechanical adjustments like HVAC, also are part of the data security ecosystem.

“AI is fueled by data, so as we start to bring in these very large data sets, and we start to establish these complex integrations, you have the potential to start exposing sensitive tenant data, financial information and more,” said Keller. “AI is transforming real estate operations, but, at the same time, it introduces this brand-new set of cyber security risks, because we’re utilizing this data in ways that we haven’t necessarily done before and it’s being done at the speed of AI.

We’re able to go through these massive data sets more quickly than we were before utilizing AI, but the concern then is that companies potentially start to over-rely on automated tools without human oversight.”

Real estate has long confronted fragmented data, limited transparency and slow decision-making. AI can help, Matías Recchia, CEO at Manhattan-based adaptive AI real estate platform Keyway, said in an email.

“It enables continuous anomaly detection, automated risk flagging and faster incident response across complex data environments,” Recchia said. “As real estate becomes increasingly data-driven, these capabilities help firms maintain tighter control over access, permissions and data integrity.

“AI allows real estate firms to move from reactive security to proactive risk management. When data flows are continuously monitored and structured, organizations gain visibility that simply wasn’t possible with manual processes.”

However, trust in AI remains a sticking point with many traditional real estate companies. That is according to a February AI adoption survey conducted by Keyway in partnership with The Appraisal, a Substack newsletter by Nima Wedlake, an investor focused on real estate technology at generalist venture capital firm Thomvest Ventures.

The survey found that trust remains the primary barrier to deeper AI adoption in high-value workflows, with 44 percent of firms saying their investment committees distrust AI-generated analysis, and only 27 percent expressing any level of trust in AI for financial underwriting. The biggest concerns include unreliable outputs or hallucinations (41 percent), integration challenges with existing systems (33 percent), and ongoing questions around data privacy and quality.

For AI to move upstream into underwriting, valuation and investment decision-making, the industry must prioritize explainability, verification workflows and fully traceable data sources, the survey found.

“Trust is the gating factor for AI in real estate,” Recchia added. “Firms aren’t just evaluating model accuracy — they’re evaluating whether outputs can be explained, verified and defended in an investment committee setting. The next wave of adoption will be driven by systems that make AI transparent, not just powerful.

“The conversation is shifting from ‘How do we protect data?’ to ‘How do we build systems where trust is embedded by design?’ In real estate, where decisions carry significant financial consequences, that shift is foundational.”

Mark Baars is a cybersecurity expert specialising in data-driven technology strategies and enterprise SaaS security awareness at the Netherlands-based Unit4, an enterprise solutions company for service organizations. He said in a statement that, “AI is rapidly changing enterprise software economics.”

Baars, who advises organizations on integrating tools while maintaining cybersecurity and operational trust, continued: “Traditional licensing and service models are being challenged, which can affect roadmap delivery, support continuity and financial stability.”

He added that organizations must ask vendors certain questions. Are the vendors financially resilient enough to withstand AI-driven shifts? Can they govern AI tools to ensure automation is safe, auditable and compliant? And, crucially, can they retrieve or migrate client data if business models or vendor strategy changes?​

The good news, said Baar, is that mature enterprise resource planning providers are already embedding AI responsibly, testing governance protocols, and planning for scenario resilience.​

MRI’s Keller emphasized that real estate companies are no different from other private and public organizations with multiple vendors and complex AI and data interactions.

“As a vendor who’s bringing AI to bear within our platform and to market for our clients, we think about the ways that we train this data, the LLMs, the way that we expose that data,” said Keller, referring to large language models. “You can’t have a massive agent that does everything. You have to have very specific agents that perform very specific tasks against a very specific set of data. And then you can start to orchestrate and string those things together.”

Philip Russo can be reached at prusso@commercialobserver.com.