AI won’t win property’s data race - Intelligence will

As generative AI becomes universally accessible, commercial real estate’s competitive advantage is shifting from the technology itself to proprietary data, relationships and human judgement.

• AI access is becoming universal, but fragmented and outdated property data can still produce confidently wrong investment answers.
• Galetti’s ReBase connects intelligence across more than 68,000 commercial properties, 153,000 legal entities and 55,000 contacts nationally.
• The next competitive advantage may belong to property firms combining proprietary data, AI speed and experienced human judgement.

AI is changing commercial real estate, but access is no longer the advantage

‍Artificial intelligence has rapidly moved from an emerging technology discussion to an everyday business tool across commercial real estate.

Over the past few years, much of the industry's attention has focused on who is using AI, which platforms they are integrating and which functions can be automated.

The arrival of generative AI platforms such as ChatGPT and Claude has changed that equation. The tools themselves are increasingly accessible to everyone. A commercial property brokerage, asset manager, developer or investor can use AI to interrogate information, summarise documents, identify patterns, analyse opportunities and accelerate research.

The competitive question is therefore shifting from “Do you have AI?” to “What intelligence are you giving it?”

For commercial property, where data is often fragmented, incomplete, inconsistent or embedded in personal relationships, that distinction could be decisive.

‍

“Interestingly, AI is already changing how prospective clients find commercial property businesses,” says Kate De Wet, Head of Marketing at Galetti Corporate Real Estate.

‍

In 2026, ChatGPT referred more traffic to Galetti's website than some social media channels that have traditionally ranked among the firm's strongest referral sources. While absolute volumes remain modest, De Wet says the direction of travel is significant.

‍

“A buyer researching industrial stock in Gauteng, or an offshore investor screening South African retail opportunities, may increasingly begin with a language model rather than a traditional search engine,” she says.

‍

“This means what these models know about a company, its listings and assets is becoming a commercial question rather than simply a technical one.”

‍

The danger of the confident wrong answer

‍The ability to process information faster does not necessarily mean the information itself is better. According to John Jack, CEO of Galetti Corporate Real Estate, this is where the real AI opportunity and risk, is emerging.

‍

“In 2026, every real estate company has access to AI. However, not every real estate company has the data to make it useful.”

‍

Commercial property presents a particular challenge because its data ecosystem is highly fragmented. Ownership information may sit in one database, company information in another, while leasing, vacancy and tenant information can reside somewhere completely different. Some of the industry's most valuable intelligence may never have entered a formal database at all.

‍

“Property data is fragmented in a way that many industries are not,” says Jack. “Ownership records sit in one place, company information in another, while leasing and tenant details may be stored somewhere else entirely.” And then there is the information held by people. “In a relationship-first industry like real estate, some of the most valuable intelligence may not be formally recorded at all. It is in a dealmaker's head or personal spreadsheet, and is also non-uniform.”

‍

The implication for investors is important: AI can analyse poor information extremely efficiently. If ownership records, vacancies, leases, decision-makers or transaction information are outdated or incomplete, sophisticated technology can still produce the wrong conclusion, only faster and with greater confidence.

AI is only as useful as the information underneath it

‍Galetti identified the value of structured commercial property information well before the current generative AI boom.

In 2014, the company began developing ReBase, its proprietary commercial real estate database, based on the premise that property ownership should be searchable and market relationships and trends more visible.

‍

“The result was ReBase, a geo-mapped platform that brings together four layers of intelligence: land data, details on the property itself, any vacancies, the commercial entity behind it and the people connected to that entity,” explains De Wet.

‍

Today, Galetti says ReBase contains proprietary records relating to more than 68,000 commercial properties, 153,000 legal entities, 371 portfolios and 55,000 contacts nationally.

The largest concentration is in Johannesburg, South Africa's biggest commercial property market, followed by Cape Town and KwaZulu-Natal and surrounding areas. But volume alone isn't the objective.

For data to create commercial value, it needs to be accurate, current, connected, interpretable and usable. That is particularly important when AI is placed on top of it. Technology is also becoming increasingly important in how property opportunities reach potential buyers.

Galetti says more than R140 million in property transactions were sold through leads generated via its social media channels during 2026, supported by increasingly sophisticated targeting designed to match properties with relevant audiences.

The combination of data, technology and distribution is therefore beginning to influence not only how commercial property is analysed, but also how buyers find opportunities and how sellers reach capital.

The data you simply can’t buy

‍There is an interesting paradox in commercial property's AI transformation. One of the most valuable technology investments a property business can make may still be the decidedly old-fashioned process of building its own database over many years.

‍

“It’s easy enough to access property data if you’re willing to pay for it. But the real value is in what you build over time and cannot shortcut,” says Jack.

‍

“We’ve spent more than a decade building ReBase and invested millions in the underlying data. The AI tools sitting on top have accelerated what we can do with it, but they aren't where the value started.”

‍

The distinction is between commodity information and proprietary intelligence. Knowing the registered owner of a commercial building is useful.

Knowing which other properties that owner controls, how those assets fit into a wider portfolio, who makes the investment decisions, where vacancies are emerging and what may be happening inside the business creates a considerably deeper layer of intelligence.

De Wet says ReBase has been built to make those connections. Galetti's proprietary information links individual properties to the companies controlling them, those entities to wider portfolios and, importantly, the people involved in making decisions.

Creating that network required more than technology. It required years of transactions, research, relationships and thousands of conversations. And that is considerably harder for a competitor or an AI model, simply to replicate.

Human intelligence becomes more valuable, not less

‍This leads to perhaps the most interesting consequence of commercial property's AI race.

The better AI becomes, the more valuable good human judgement may become alongside it.

‍

Jack uses a simple example. “Consider two identical data points. A company owns a particular building and has held it for ten years. Any decent real estate database can tell you that much.”

‍

But the broker who knows the owner may know something more. The company could be consolidating its offices. The building may no longer be strategically important. Another investor could be assembling properties in the surrounding precinct. A major tenant may be considering relocating.

‍“The data is the same; the context changes what you do with it,” says Jack. For investors, landlords and developers, this distinction matters. AI may be exceptionally good at searching large datasets, identifying correlations and surfacing connections that would take a human analyst considerably longer to uncover.

But identifying a connection and understanding its commercial significance are not necessarily the same thing. That still requires market experience, local knowledge, relationships and judgement.

A different question for property investors

‍For property owners, buyers and sellers choosing advisers and service providers, the AI conversation may therefore need to change.Instead of simply asking: What AI tools do you use?

The more valuable questions could be: What data are those tools running on? How current is that information? What proprietary intelligence do you have that others don't? And who interrogates the AI's output before it influences an investment decision?‍

Jack believes this is ultimately where AI creates the greatest value.
‍

“The purpose of these tools is to make good people better at what they already do well.”

‍

“They can surface connections and interrogate information faster than any analyst can. But it still takes someone who understands the market to recognise which connections actually matter.” For commercial property investors, that may be the central takeaway.

‍

The industry's AI race is unlikely to be won by whoever has the smartest chatbot. It will be won by those who combine better data, deeper market intelligence, stronger relationships and experienced human judgement and then use AI to make all four work faster.

‍