Here’s what’s on our plate today:

• 🏗 AI is reshaping real estate—from valuations to construction.

• 🧠 Ellison’s TikTok obsession, OpenAI’s tangled empire, and humanoid robots.

• 📊 How would you feel using AI to guide big-money decisions?

• ✍️ Turn messy property data into smart predictions using GPT-4 or Claude.

Let’s dive in. No floaties needed…

The Laboratory

Can AI models streamline real estate purchases?

Even before entering politics, the current president of the United States had built his reputation in real estate,  a notoriously unforgiving industry where success depends on navigating land deals, politics, corruption risks, and even brushes with organized crime. His ability to master those challenges not only added towers to the New York skyline but also helped cement his public image as a dealmaker.

He is hardly alone. Across the world, real estate has been a proving ground for ambitious figures willing to take on its labyrinth of permits, zoning approvals, and volatile markets. But while the industry has produced larger-than-life personalities, it has been much slower to embrace innovation. In many places, brokers still rely on the same old methods like market reports, local comparisons, and intuition to estimate a property’s worth.

That may be about to change. Like many industries before it, real estate could be on the cusp of a major transformation, driven by the data-crunching power of artificial intelligence.

AI’s value proposition in real estate

Deep learning is very good at spotting patterns in data sets, and the real estate industry is data-rich. This makes the real estate industry an ideal candidate for testing and deploying AI models.

One of the areas where AI can reshape the real estate industry is by removing redundancies and improving efficiency.  According to Morgan Stanley, AI can automate 37% of tasks in real estate, representing $34 billion in operating efficiencies. Even now, companies that report using AI for on-site staffing reported higher satisfaction from clients and their teams.

Other than staffing, Morgan Stanley predicts that AI  would be used in areas like management, sales, and related activities; office and administrative support; and installation, maintenance, and repairs. This, it said, could result in $34 billion in efficiency gains for the real estate industry by 2030.

The gains are primarily expected to be achieved by using machine-learning models to digest years of sales history, tax records, images, and structured data to produce automated valuations. This would make price prediction faster, cheaper, and more scalable for brokers, buyers, and lenders.

Beyond data crunching, search and discovery are also expected to get a major overhaul with AI lowering costs and accelerating content creation.

Easing the hunt

Property descriptions, virtual staging images, 3D tours, and even floor plans can be produced automatically using machine learning. This would save brokers by reducing costs associated with photography and staging budgets.

Matterport, a company that specializes in digital twins of physical spaces, has expanded its AI-powered offerings and was acquired for $1.6 billion by CoStar, highlighting the rise in interest in AI-powered solutions.

Another area that has historically been a pain point for developers and buyers has been document processing. Even here, AI can be used to extract information from pay stubs, contracts, and disclosures in seconds, flagging issues in due diligence that once required hours of manual review. For borrowers, AI can also streamline the process of applying for a mortgage, accelerating loan decisions and broadening access to financing.

Cutting construction costs

In construction, AI and computer vision are automating quality control and project tracking. Site photos, drones, and 360-degree cameras are used to detect defects, measure progress, and compare work completed against the plan. Companies like Doxel, OpenSpace, Track3D, and Procore are rolling out AI tools that forecast delays and help contractors catch problems before they balloon into costly overruns.

For larger investors, AI is being deployed to aggregate data like event trends, macroeconomic indicators, and even foot traffic to assist them in strategising their acquisitions, pricing, and portfolio rebalancing.

The real-world use cases of AI in real estate have also allowed startups to carve out a space for themselves.

Track3D, a startup specializing in AI-first solutions, recently raised $10 million in Series A funding led by Ironspring Ventures. The company is working on purpose-built solutions aimed at modernizing construction monitoring.

In 2024, Reuters reported that CoStar Group would buy Matterport in a cash-and-stock deal valued at $1.6 billion. The deal would allow the real estate information provider to win listings and sell more properties by enhancing 3D virtual tours. It was also reported to give CoStar tools to automate design suggestions, interior layout optimization, and richer property presentation.

However, while AI is being deployed to streamline processes and cut operational costs, AI models are not infallible, and when it comes to the real estate sector, biases and miscalculations can lead to devastating losses.

The risks

In what was a warning call for automating too much too fast, in 2021, Zillow, an online real estate marketplace, lost over $300 million, forcing its shutdown and layoff of 2,000 employees.

The company had been using machine learning for listings; however, seeing their success in predictive tech, it began using algorithms to purchase and flip homes. The model used machine learning on historical sales, tax records, and home features.

However, while the algorithm worked well in normal conditions, when the housing market heated up in 2020–21, prices moved too quickly for the model to keep up. This resulted in Zillow overpaying for thousands of homes because its algorithm assumed appreciation would continue steadily. By the time the houses hit the resale market, conditions had shifted.

Similarly, while AI can help screen job listings, overreliance can be problematic, as was seen with Amazon in 2018. According to a Reuters report, Amazon was using a recruiting engine that was trained to vet job applications by observing patterns in resumes submitted to the company over 10 years, most of which came from men. This led to the system favoring male candidates over females. While Amazon did rectify the problem, there is no guarantee that machines will not devise other discriminatory ways of sorting candidates due to historical biases.

In real estate, this could result in problems not just in hiring, but also in amplifying the problem of segregation. AI models trained on historical data may not be able to overcome age-old biases and could further profile buyers based on race, educational background, and whatever datasets the algorithm is fed.

Another problem is authenticating the paperwork. Generative AI can be leveraged to fabricate audio, video, documents, or reviews that appear legitimate. In real estate, fraudsters have used AI to clone voices, send fake wire-change requests, or present deepfake ‘sellers’ in video calls. In one case from California,  a family reportedly lost $720,000 to a voice-imitation scam.

AI can, on one hand, bring down the cost of construction while improving accessibility and quality construction, but it can also amplify existing problems and introduce new ones.

AI and real estate

Real estate has long been one of the key pillars of economic growth, often dominated by those with the resources and resilience to navigate its complexities. AI now offers a chance to open that space, making the industry more efficient, transparent, and inclusive.

But the lessons of the past are clear: algorithms are powerful, not infallible. Overreliance on them can reinforce old biases, concentrate power, or even enable fraud. The true potential of AI lies not in replacing human judgment, but in amplifying it by streamlining transactions, improving construction, and widening access without losing sight of the people behind every deal.

If used wisely, AI could help real estate shape not only today’s markets but also the opportunities of generations to come.

Roko Pro Tip

Prompt Of The Day

Bite-Sized Brains

• Ellison’s TikTok play: Oracle’s Larry Ellison is angling for a starring role in global tech power, and TikTok might be his ticket.

• OpenAI’s deal web: From Microsoft to Nvidia to Apple, OpenAI’s overlapping partnerships are raising questions about influence and independence.

• Robot bubble warning: A top roboticist says the humanoid robot hype is unsustainable — and a collapse is coming.

Tuesday Poll

Rate This Edition