Surviving the "SaaSpocalypse": A strategic guide to real estate AI

The economics of software have shifted. Because artificial intelligence can generate working code quickly and cheaply, corporate real estate (CRE) and technology leaders are asking a reasonable question: do we still need to pay for PropTech licenses, or could we just have AI build what we need instead? 

For real estate investors and occupiers, the pressure is real. Capital costs keep climbing, as does the criteria for what counts as a “modern” portfolio. (Things like smart-building systems and ESG tracking tools require continuous investment.) Seventy-eight percent of business and commercial real estate (CRE) leaders say AI will significantly reshape their portfolio strategy over the next three to five years, according to JLL's 2026 Future of Work Survey. But only 15% have gotten past pilots into real, operational use.  

Some industry analysts have connected the relative ease of creating code with AI and the recent fall in software market capitalizations and deemed the event, the “SaaSpocalypse.” But the reality is, not all PropTech software can be built from scratch. 

That gap between what you believe and what you’ve built, is exactly where expensive mistakes happen. (Read our full white paper here.) 

Here's the short answer: it depends on the workflow. We sort every decision into one of three lanes:

Andy Targell, Global Head of Real Estate Technology Advisory at JLL Technologies (JLLT), explains why it’s vital to stay in the right lane when it comes to your tech spend:

"Generating software is not the same as having a team maintain it for years afterwards," he says. “Using a powerful AI tool to rebuild standard corporate systems like lease accounting, which must comply with strict accounting standards, is a poor risk-adjusted use of capital, especially when software represents only a fraction of total enterprise real estate spend.”

According to MIT’s 2026 research into AI in business, AI initiatives built entirely in-house succeed only about a third of the time. Analysts at Gartner, McKinsey and others have been tracking this subject for some time and the pattern breaks down into three forces:

This isn't theoretical: in one widely reported case, an AI coding agent deleted a live production database during an active code freeze, then misreported what had happened. The cause was a bad safety net, with no approval gates and no separation between test and live environments. 

Keep in mind, the risk isn't the AI; it's whether you've built the guardrails around it.

It's also worth remembering that your SaaS vendors are cutting their own costs with AI without always passing the savings to you. That’s one more reason to know exactly what a “build” would really save you before you commit to it.

For the first time in 15 years of JLL Future of Work research, skills and capability gaps have overtaken budget as the number one constraint on real estate transformation. Money alone doesn't solve this problem. Without people who know how to select the right platform, drive adoption and prove it's working, even a well-funded team can stall.

Four things separate the projects that make it from the ones that stall:

• Clean data: Fragmented building names or unstandardized lease terms get amplified by AI, not fixed by it.
• Simplified architecture: Most large companies run hundreds of disconnected real estate, finance and HR applications, but only a handful are properly connected.
• Clear ownership: When anyone can build an app in an afternoon and point it at live portfolio data with no review and no owner, don’t be surprised when it breaks something.
• Risk controls: Security scanning, a separate test environment and a human who signs off before anything touches your corporate network.
   

That's the real tension: the same technology that boosts productivity also opens new risks. In our Future of Work survey, CRE and C-suite leaders ranked three of their top four portfolio risks as technology-related: cybersecurity and data privacy (47%), tech and AI disruption (41%) and uncertainty about how AI will change space needs (40%). Your job is to make sure each investment reduces risk more than it adds.

A few questions we hear constantly from occupiers weighing this exact decision:
 

Why do in-house PropTech builds fail more often than partner-delivered projects?
 

MIT’s research on AI in business found in-house builds fail about twice as often as partner-delivered ones, mainly because most organizations haven’t built the clean data, integration and change-management foundations that experienced partners already have.

Should we replace our existing IWMS/Lease Administration software with custom AI agents?
 

No. Your core systems of record are the safest, most cost-effective default. AI works best when it "boosts" such platforms through integrations, not when it replaces them.

How do we know if we should "build" or "buy" a new real estate tool?
 

Buy when the off-the-shelf platform already covers the job, like standard work order ticketing. Only build when the workflow is genuinely unique to you, gives you a real edge and depends on data no outside vendor could ever replicate.

Not sure whether your next tech decision is a buy, a boost or a build? JLL Tech Advisory will walk your stack against this framework with you, workflow by workflow, and show you exactly where the risk sits.