AI has Changed the value of proptech and legacy technology is paying the price
Thought leadership by Andrew Stanton
For more than two decades, the value of proptech was built around a fairly simple proposition: take an inefficient property process, turn it into software, sell it to an enterprise customer and charge a recurring subscription. That model created some very valuable companies.
If we focus on for example the tiny world of residential estate agency/brokerage here in the UK, we see that property portals, CRM platforms, property management systems, valuation platforms, transaction software and data businesses built substantial recurring revenues on the back of technology that was, by the standards of today, relatively unsophisticated. Their competitive advantage came from being embedded in customer workflows, owning valuable datasets and, importantly, becoming difficult to replace.
AI is now challenging that equation. The issue is not simply that artificial intelligence makes software better; the bigger issue is that it is forcing investors, acquirers and customers to reconsider where the value actually sits inside a technology business. For some established proptech companies, that could be uncomfortable. The technology itself may no longer be the asset the market once assumed it was.
The problem with the old software model
Traditional proptech was largely built around interfaces. Users logged in, navigated dashboards, entered information into forms, ran reports, moved between screens and learned how the software worked. The software was the interface through which the customer accessed the underlying capability, and over time that familiarity became part of the product’s defensibility.
AI changes that relationship. Increasingly, the user does not need to know which screen to open or which report to run; they can simply ask the system what they need, and increasingly AI can retrieve the information, interpret it and take action. That is a profound change. If an AI agent can interrogate several systems, extract the relevant information and produce the required outcome, the value of any individual application sitting underneath it potentially falls.
The question becomes: why am I paying for ten pieces of software when an intelligent system can orchestrate the information and functionality I need across all of them? This is precisely the structural threat facing traditional SaaS. The issue is not necessarily that every established application disappears, but that the market may begin to view many of those applications as interchangeable components rather than strategically important technology platforms.
Proptech is particularly exposed
Property technology may be more vulnerable than many other software categories because so much of it was built around information management and workflow automation. Think about the traditional proptech stack: CRM, property management, valuation, listings, market intelligence, document management, lead management, transaction management, maintenance and analytics. Each solved a particular problem, but AI is increasingly capable of sitting across those workflows rather than simply operating inside one of them.
The emerging model looks less like a collection of applications and more like an intelligent operating layer sitting above them. AI can increasingly connect information from multiple systems, understand the context and deliver an outcome without the user necessarily knowing which underlying application provided the information. PwC and ULI have already described the emergence of a potential “property operating system” built around AI agents, digital twins and integrated data layers operating above — and potentially eventually replacing parts of — legacy platforms.
That is potentially much more significant than simply adding a chatbot to a property management system. A chatbot improves an existing product; an AI operating layer potentially changes the architecture of the entire market. The distinction matters because the former protects the incumbent while the latter can undermine the assumptions on which the incumbent’s valuation was built.
The valuation problem
This creates a particularly difficult problem for established proptech businesses. For years, investors valued software companies on recurring revenue, customer retention, growth and margins, and those metrics still matter. But increasingly, investors are asking a different question: how defensible is the software when AI changes the cost of creating functionality?
A company with £20m of recurring revenue may look attractive if customers are deeply embedded in its platform and switching costs are high. But what happens if an AI-native competitor can deliver 80% of the functionality at a fraction of the cost? Or if a major customer decides it can build an internal AI layer that sits across several existing systems? Or if the customer stops interacting directly with the software altogether?
The risk is not necessarily that revenues disappear overnight. The risk is that the multiple applied to those revenues changes. That distinction is critical: a legacy proptech business can remain profitable, cash-generative and operationally successful while simultaneously becoming worth less because the market no longer believes its technology provides the same degree of future competitive advantage.
Not all legacy proptech is doomed
This is where the argument becomes more interesting. AI does not automatically make old technology worthless; in fact, some established proptech companies may become more valuable precisely because of what they have accumulated over the last 10 or 20 years. The real asset may not be the software itself, but the data, customers, integrations, workflow position, relationships and institutional knowledge sitting behind it.
Some of the strongest established platforms have spent decades becoming deeply embedded in the property industry. They may hold proprietary datasets, have thousands of customers, connect to hundreds of other systems and sit directly inside critical business processes. Those assets are considerably harder to reproduce than a software interface.
This creates two very different categories of legacy proptech. The first is software whose primary advantage is functionality, and that is vulnerable because AI can increasingly reproduce functionality. The second is software that has become infrastructure for an industry — deeply embedded, connected to proprietary data and sitting at the centre of important workflows — and that is considerably harder to displace.
The uncomfortable middle
The most interesting companies may therefore be those sitting between the two. They have established customers, valuable datasets and years of accumulated property information, but underneath the hood they may still be running technology architectures designed for a very different era. Adding an AI assistant does not necessarily solve that problem.
The question is whether the company can turn its accumulated assets into an AI-native proposition. That requires more than putting “AI-powered” on the website; it means rethinking the product, pricing model, user interface and ultimately the relationship with the customer.
The real opportunity for established proptech may therefore be to use AI to unlock assets that were previously trapped inside legacy systems. A 15-year-old property database may suddenly become dramatically more useful when an AI system can interpret it, connect it with external information and turn it into actionable intelligence. The technology may be old, but the underlying data and relationships may be extraordinarily valuable.
The next valuation divide
The proptech market is therefore beginning to develop a new valuation divide. On one side are companies whose value is primarily derived from software functionality that AI can increasingly reproduce. On the other are companies whose value sits deeper — in proprietary data, distribution, relationships, integrations, workflow control and industry-specific intelligence.
The first group faces multiple compression. The second may find that AI actually increases the value of what they have spent decades building. The irony is that the companies with the oldest technology may sometimes possess the most valuable raw materials for the next generation of technology.
That is why the next phase of proptech will not simply be about who has the best AI. It will be about who owns the most valuable position from which AI can operate. And that may turn out to be the most important distinction in determining which of today’s established proptech companies become tomorrow’s infrastructure — and which become yesterday’s software.
