Clean data, smart decisions: Implementing AI successfully in CRE

Listen to this article

By Dave DuVarney
principal with Baker Tilly‘s digital solutions practice

Artificial Intelligence is rapidly changing the business landscape, and the commercial real estate industry is starting to accelerate its adoption of AI. Historically a laggard when it comes to technology, commercial real estate (CRE) has a history of being a “deal house” industry focused on moving fast, closing deals and worrying about the data later. The statistics are telling: The real estate industry giant JLL reports 87% of companies are increasing their real estate tech budgets, but 60% are unprepared for strategic AI integration. As CRE looks at the benefits of adopting AI, the industry is simultaneously being forced to slow down and look closely at the frequently overlooked area of data.

The problem with dirty data

AI adoption issues often stem from data issues. Quite simply, bad data equals bad decisions. Part of the problem is that there is no common definition among the terms being used by employees. A lack of shared business language among CRE teams often results in a decision-making breakdown. For example, if you ask what your non-operating income (NOI) has been across metro-Milwaukee over the last 12 months, you must have clearly identified properties, clear geographic designation, solid NOI figures, and a consistent understanding of them. If any of those definitions are fuzzy, the answer is unreliable.

Data can also break down across system boundaries. In real estate, when a property moves through different systems (i.e., investment management, asset management, and property management), tracking can become a challenge. Maintaining data consistency throughout these transitions is not automatic.

People also assume AI can fix dirty data. The opposite is true; AI just makes the problem more visible. While AI can spot anomalies and flag it for a human, it cannot find the missing data itself and cannot make it up.

First foundational step: Clean data

Effective use of AI starts with clean data, and clean data starts with building a solid data foundation. CRE companies need to establish data governance and an operating model to support the governance implementation. Any incoming data needs to be sourced and aggregated in a consistent, usable way.

As internal data input becomes more clean, consistent and reliable, companies can start integrating data with external sources like market data for a more holistic analysis. On the other hand, if companies do not create accountability for data in the governance structure on the input side, the results are going to get messed up and AI will not magically fix it.

Where AI adds value in CRE

AI adds value to real estate companies in three main areas: data-related workflows, investment modeling and property and contract management. Most AI adoption in CRE is on the back-office side where AI is changing the equation by making previously cost-prohibitive tasks like data aggregation and financial analysis more feasible.

AI and predictive analytics are also being used in investment risk modeling, including the assessment of market data, prescreening of deals and automated property valuation models. AI can also be used in the due diligence process as the technology can analyze vast amounts of data integrated from different data sources. However, people should always be part of the process. AI is a tool that informs and accelerates, but it cannot replace human accountability or domain knowledge. AI can spot a problem or an anomaly like a missing rent roll, but a human still needs to fix it. We call this “human in the loop.”

Opportunity and urgency in a competitive landscape

With a thoughtful and deliberate start to their AI exploration, CRE firms will also have a competitive advantage. But the timing and implementation sequence matter: governance, aggregation and data quality should always come first. Companies should not chase AI use cases and pilot projects unless they have the infrastructure in place to support it.

If CRE organizations have already started their AI journey with bad data and are hitting a wall, they need to press pause, go back to the business processes and address the source of the problem. Businesses also need to realize that corrections are iterative, not a one-time project.

Indeed, CRE firms that are best positioned for success think about all the dimensions of AI readiness before jumping in. This means identifying specific uses for AI and highlights the contributions an AI investment can bring before implementation. Data management includes feeding “clean” data into an AI model to train the model and expand its capabilities.

As with any digital initiative, the IT environment must include strong cybersecurity controls. Risk, privacy and governance should address the financial, regulatory and reputational risks that could affect the organization. Another key consideration for CRE is AI adoption, including the company’s openness to change, employees’ data and AI literacy and potential barriers to successful implementation.

As AI reshapes commercial real estate, success hinges on thoughtful, deliberate adoption. Organizations need to prioritize clean, well-governed data as an initial step because even the most advanced AI is only as reliable as the foundation it’s built on.