Interacting with AI has become a normal part of professional life in a way that would have seemed remarkable just three years ago. Drafting emails, summarizing documents, answering research questions, generating first drafts, these are tasks that millions of people now hand to ChatGPT or Claude without much thought. Commercial real estate has followed that curve. According to Kolena’s 2026 State of AI in CRE report, 34% of CRE firms are now using a general-purpose AI tool as part of their regular workflow. The number sounds encouraging until you read what sits next to it: 78% of those same firms still process their documents manually. Nearly a third of the market has a general AI subscription and a document workflow that is entirely, stubbornly, manual.

That gap is the defining feature of where commercial real estate’s AI adoption actually stands, and it points to a distinction that the industry is beginning to understand more clearly. Conversational AI and automated AI are different tools built for different purposes, and using one doesn’t put you any closer to the other. “ChatGPT and Claude are great for one-off tasks but now people realize that those models are not for large-scale automation,” said Mohamed Elgendy, CEO and co-founder of Kolena, which builds purpose-built AI agents for document-heavy workflows in commercial real estate, lending, and insurance. “If you need to process a large amount of documents, you don’t have time to talk to AI about each task, you just need it to do it.” Companies using a general LLM are barely further along in production deployment than those using none. Without thoughtful deployment of AI at scale the adoption is plateauing for many organizations.

The companies that have moved past that plateau are the ones that have committed to purpose-built automation rather than general-purpose tools. That shift has required CRE organizations to overcome the risk of trusting consequential work to a model that can produce convincing but incorrect outputs. In real estate, where lease terms, financial figures, and property data carry direct business and legal weight, an undetected error isn’t a minor inconvenience. It is a real liability. 

Despite the concerns, there are ways to create safeguards against the occasional mistakes made by AI. “Our agents are always looking over the shoulder of our clients’ agents so any hallucinations get caught,” Elgendy said. “When it catches a mistake, it will then reengineer the prompts in a way that will prevent any misunderstanding.” That supervisory layer, where the system monitors and corrects its own outputs, represents a more mature approach to AI deployment than most organizations encountered in their early experiments with general-purpose tools.

One of the more surprising discoveries for organizations moving into serious AI deployment is that the efficiency gains are only part of what they are getting. “There is a hidden value to process automation that most teams have not considered,” Elgendy said. “It is able to document how the processes are getting done and update them across the organization if regulations or internal policies change.” When an AI system executes a workflow, it creates a continuous record of exactly how that workflow was performed. That record becomes a living document of the organization’s actual operating procedures, which is a fundamentally different artifact from the documentation that most organizations produce through periodic manual review.

Recording processes are particularly important for real estate companies operating across multiple states or countries, where the same underlying process may need to adapt to different regulatory frameworks in different markets. Historically, mapping those variations required sending analysts into each market to interview staff, compile what they found, and produce documentation that became outdated the moment anything changed. The process took months, the outputs went stale quickly, and the result was a set of process documents that described an idealized version of how work was supposed to happen rather than an accurate picture of how it was actually happening. An AI-automated workflow doesn’t just perform the task. It continuously updates the organization’s understanding of how the task is being performed and can flag when actual execution diverges from policy or when a regulatory change in one jurisdiction requires an adjustment that needs to propagate across the rest of the organization.

Most companies haven’t fully accounted for the fact that process automation generates organizational intelligence as a byproduct. “Companies are learning about themselves because sometimes there is a skew between how executives think things are being done and how the work is actually being completed,” Elgendy said. The gap between intended process and actual process is one of the most persistent problems in management, and it tends to be invisible precisely because the people who know how work is actually done are the people doing it rather than the people overseeing it. AI that executes and documents workflows surfaces that gap in a way that periodic audits and management interviews rarely do.

Business continuity is another downstream benefit that gets least attention in the AI adoption conversation but may ultimately prove most significant. Institutional knowledge in most real estate organizations lives in people, in the experienced analyst who knows how to handle a non-standard rent roll, the asset manager who has the mental model for a particular portfolio’s quirks, the compliance officer who knows which state regulations require extra attention. When those people leave, that knowledge leaves with them. A workflow that has been automated and documented doesn’t have that problem. The knowledge is embedded in the system, not the individual, which means it survives turnover in a way that traditional knowledge transfer approaches rarely achieve. For an industry where experienced talent is expensive and competitive, that continuity value deserves to be part of how organizations build the business case for AI investment, alongside the more obvious argument about hours saved and tasks automated.

The Kolena report’s most pointed finding may be its simplest. Of every commercial real estate company with an open question about AI adoption, 87% have only internal work left to do. Just 13% face a genuine hard blocker in the technology itself. The most common obstacle, cited by more than half of those companies, is simply agreeing internally on what to use AI for in the first place. The technology is ready. The use cases are proven. The economics are clear enough that more than half of the companies in the study described themselves as being in high or critical pain around their document workflows. What is holding back the majority of the industry is not a technical problem. It is the organizational work of deciding what to commit to, building the internal alignment to act on that decision, and accepting that AI’s value in commercial real estate is no longer theoretical. The companies that have already made that decision are not just saving time. They are building a more legible, more resilient, and better-documented version of their own organizations. That turns out to be worth considerably more than the hours they are getting back.

The post The Unexpected Ways AI Is Changing How Commercial Real Estate Organizations Operate appeared first on Propmodo.

​