The interesting thing about Blueprint Las Vegas 2026 wasn’t how much the agenda talked about AI. It was how often it moved past it. Across conversations about operations, maintenance, brokerage platforms, and enterprise technology, the same question kept coming up: does this technology create value once it leaves the pilot?
Our team spent September 22–24 at Blueprint, and one topic came up in conversation after conversation: data. It surfaced across very different parts of the market, most often in the context of AI. The questions were rarely about which model to use. They were about whether the data underneath is clean, connected, and accessible enough for AI to produce something a team can act on.
That question is the gap this article is about. Adoption is putting new tools into use, running pilots, and adding AI capabilities. Value is connected data, workflows that actually change, measurable outcomes, and products that hold up in production.
Here is what we think closes that gap, from the perspective of a team that builds real estate software every day.
The Pilot Is No Longer the Finish Line
AI conversations at Blueprint were less about what the technology could theoretically do and more about where it can create value in real operations. The examples spanned development, construction, property operations and leasing, with companies exploring AI for tasks ranging from reducing delinquency and auditing leases to capturing institutional knowledge. Just as importantly, the conversations acknowledged that there is no single path to adoption: some companies are buying off-the-shelf tools, others are building or co-building solutions, and some are choosing not to use AI where the business case isn’t clear.
For us, that gap makes sense. In real estate projects, the hardest part is rarely adding AI. It is making the surrounding system reliable enough for AI to matter.
For AI to survive beyond the demo, five layers have to work together:
1.Trusted data, with known gaps.
2.Connected systems, so outputs land where people already work.
3.Workflow ownership, with a named person and a decision the output supports.
4.Traceable outputs, so results can be checked against their source.
5.Human oversight where a wrong answer is expensive.
AI is the visible layer. The infrastructure underneath decides whether it works.
Technology Has to Earn Its Place In the Stack
The question is no longer simply “should we adopt it?” but “does it earn its place?” Across conversations at Blueprint, technology was increasingly discussed in terms of what it changes in day-to-day operations — where it reduces friction, saves time, improves decision-making, or replaces work that teams are still doing manually.
That focus reflects a broader shift in the market. Deloitte’s 2027 outlook points to greater selectivity in technology and capital spending as CRE firms evaluate where new capabilities can deliver measurable value. JLL’s mid-year outlook also identifies cost management as a primary decision driver for corporate real estate leaders.
This changes what a good product brief looks like. A feature that can’t be connected to a measurable business outcome has a harder time surviving the next budget review. Useful measures might include turn time, work-order close rate, hours spent on reporting, or NOI.
Product teams should expect a harder question than “what does it do?” They should expect “what does it replace, and how will we know it worked?”
Data and Integrations Decide Whether AI Works
In our conversations at Blueprint, data kept coming up across very different parts of the real estate technology stack. AI was often what brought the issue to the surface, but the underlying challenge was usually the same: can the right data be captured, connected, and accessed when a workflow needs it?
That applies to everything from operational data and property records to market intelligence and reporting. AI can only work with what the system can reliably provide. If information is incomplete, inconsistent or locked in separate systems, even a strong AI capability has limited room to deliver value.
The structural issue is well known. A February 2026 Propmodo analysis, citing the data-standards consortium OSCRE, describes real estate data as fragmented, inconsistently formatted and spread across disconnected systems. Many firms still rely on bespoke integrations between property management, accounting, leasing and reporting tools — and those integrations often need to be revisited as the underlying systems change.
In practice, a real estate product rarely owns its inputs. It has to reconcile property systems, CRMs, accounting platforms, listing feeds, market data and building systems, each with its own definitions and failure modes. “Rent” in one system and “base rent” in another is a small mismatch until a model aggregates it across 40 properties.
The same challenge appears in building performance. Operators may already have meter, BMS, and IoT data, but turning those signals into an actionable alert, work order, or savings opportunity still requires integration.
This is why we treat normalization, enrichment, and pipeline reliability as product features, not back-office plumbing. Data infrastructure is not plumbing when the product depends on the data.
Legacy Systems Aren’t Going Away
Digital transformation in real estate happens alongside existing systems far more often than instead of them. Companies are not starting with a clean slate. They already have enterprise platforms, property management systems, CRMs, accounting tools, and internal applications that run critical parts of the business.
The scale of that challenge is not new. Deloitte’s 2024 Commercial Real Estate Outlook found that 61% of respondents said their core technology still relied on legacy systems. Its 2027 outlook continues to identify legacy processes and fragmented data foundations among the barriers to scaling AI.
That creates a more practical modernization challenge. The goal is rarely to replace everything at once. It is to connect what already works, modernize where the gaps are, and introduce new capabilities without creating another disconnected layer.
The same applies to AI. Adding an AI capability on top of an existing system does not automatically make the underlying workflow better. Security, data access, integration, maintenance, and the cost of supporting what gets built all become part of the equation.
The accounting system that closes the books every month carries years of business rules. So does the property management system every site team knows. Ripping them out can create more risk than it removes.
A more reliable path is to connect, modernize, and replace selectively:
- Wrap the systems that work in an integration layer and clean APIs.
- Move one workflow at a time onto new components, starting where the pain and payoff are clearest.
- Replace a component outright when extending it costs more than rebuilding it.
Modernization is less about choosing between legacy and new technology and more about creating a path between them.
The Real Gap Is Between Prototype and Production
“Can we build a prototype?” is now a much cheaper question. Off-the-shelf models and AI-assisted development have made it possible to turn an idea into a convincing demo faster than before.
The harder question is whether that prototype can become a reliable product that people use across an organization or portfolio. Production introduces a different set of requirements: integrations, data quality, security, access controls, monitoring, maintenance, cost management, and clear ownership. What works in a demo can behave very differently once it becomes part of a real workflow.
Deloitte’s 2027 Commercial Real Estate Outlook puts numbers behind this gap: 92% of surveyed CRE organizations were still in the research or pilot phase for AI, while 8% reported integrated AI solutions. The same research points to data foundations, legacy processes, and governance as important constraints on moving from experimentation to scaled adoption.
The shift from prototype to production is ultimately a shift from demonstrating what technology can do to making it reliable enough for a business to depend on. That is where much of the real engineering work begins.
In our experience, the gap between a pilot and a product shows up in the same places:
- Architecture that holds up when one property becomes 400.
- Integrations that survive vendor updates and API changes.
- Data monitoring, so bad inputs are caught before they reach a report.
- Security and access control that pass an institutional owner’s review.
- Workflows designed with site teams, not only for head office.
- Maintainability, so the product can keep changing after the original team moves on.
None of this shows up in a demo. It is also most of the work. The teams that plan for it during the pilot, even lightly, reduce the risk of rebuilding the product later.
What Real Estate Technology Teams Should Do Next
1.Start with the workflow, not the AI tool. “Use AI for leasing” is not a spec. “Cut lease-audit time per property from two days to two hours” is.
2.Treat data and integrations as product infrastructure. Budget for normalization, pipelines and every system connection the same way you budget for UI.
3.Modernize around existing systems. Connect what works, move workflows in slices, and replace selectively.
4.Define measurable value before building the pilot. Pick one or two metrics the operator already tracks, and agree on what “worked” means.
5.Design for the next stage of scale. A pilot doesn’t need enterprise-grade everything, but its core architecture shouldn’t make the next stage of growth require a rebuild.
From Adoption to Value
At Blueprint, one thing became clear. The next challenge isn’t getting more technology into real estate. It’s making the technology already being adopted work together: data that can be trusted, systems that talk to each other, and workflows that change because of it.
At ORIL, we build and modernize the software, data infrastructure, and integrations that turn real estate technology from a promising pilot into something teams rely on. If you’re working through these questions — modernizing a legacy stack, connecting fragmented data, or taking an AI initiative from pilot to production — our real estate technology experts can help you build the systems behind that transition.
