The conversation about artificial intelligence in multifamily real estate has largely been framed as a question of location: which platform should own the AI layer? The property management system? The leasing CRM? The resident experience application? Alex Samoylovich, co-founder of Livly and a veteran of institutional multifamily development through CEDARst Companies and Proper., argues that this framing misidentifies the constraint.
Multifamily technology stacks did not evolve as coherent operating systems. They grew as collections of specialized tools — each capturing a slice of operational reality with its own identifiers, timestamps, exception logic, and data definitions — stitched together around a system of record that was never designed to serve as a single source of truth.
The statistical picture confirms what operators already experience. Research from NMHC and RealFoundations published in 2024 found that 78 percent of multifamily operators run ten or more applications within their technology stacks. The portfolio’s operational reality is not concentrated in one system. It is distributed across dozens.
Centralizing the AI model does not change this distribution. When a centralized AI is fed data from a fragmented stack, it receives a fragmented view of reality. It cannot reliably coordinate what it cannot reliably reconstruct.
Samoylovich’s operational background informs a specific diagnosis. Through years of developing and managing large-scale multifamily portfolios, the pattern he observed was consistent: portfolio risk and performance degradation do not originate within individual systems. They originate in the transitions between them.
Livly’s framework refers to this as the seams tax — the compounding cost of coordination failures that occur when approvals, exceptions, delays, vendor handoffs, access changes, and privileged actions like money movement, notices, and dispatch are not tracked with consistent identity, timestamps, and decision attribution across all systems involved.
The economic evidence supports this characterization. Data from TechRadar summarizing HubSpot survey findings published in October 2025 shows that 74 percent of businesses perform manual data transfers weekly as a direct response to system fragmentation. Thirty-four percent report measurable revenue loss attributable to disconnected systems. Nine percent of organizations report trusting their data for reporting purposes — meaning the overwhelming majority operate on data they do not fully trust.
The adoption curve for AI in real estate and across industries broadly is not the constraint. McKinsey’s State of AI report from May 2024 documented that 72 percent of organizations had adopted artificial intelligence, with 50 percent deploying it across two or more business functions. Adoption, as Livly’s framework observes, is mainstream.
The gap is structural. The Zapier AI Sprawl Survey published in December 2025 found that 90 percent of organizations identified orchestration as a critical unmet need, while 76 percent reported measurable harm from disconnected AI tools operating without coordination. Only 18 percent had established a formal AI governance structure.
This gap — high adoption, low governance readiness, high orchestration demand — is the condition that makes centralized AI feel powerful in demonstrations and brittle in production. The models exist. The tools exist. The portfolio-level control plane that coordinates decisions safely across systems does not yet exist at scale.
The framework that Samoylovich and Livly are advancing argues that the next evolution in multifamily AI is not another assistant embedded in another interface. It is the development of shared operational truth over time — a portfolio-wide capability for every system to agree on what happened, in what order, and why, particularly when actions carry real consequence.
This requires three foundational elements to be treated as strategic infrastructure rather than implementation details: identity consistency across all systems involved in a coordinated action, temporal integrity so that decision sequences are reconstructable after the fact, and decision receipts that create auditable attribution for privileged actions. Without these elements in place, AI cannot move from a helpful tool operating within a single application to a safe execution layer operating across a portfolio.
For Alex Samoylovich, this is not an abstract position. It reflects the operational thesis he has tested across institutional multifamily at CEDARst, standardized across geographically distributed portfolios through Proper, and is now building into the product architecture at Livly. The seams between systems are not incidental. They are where the industry’s most significant performance and risk outcomes are determined — and where AI, if properly governed, has the greatest opportunity to deliver durable value.
What is the Shared Operational Truth framework developed by Livly and Alex Samoylovich?
Why does Alex Samoylovich argue that centralized AI is insufficient for multifamily operations?
What is the 'seams tax' in multifamily property management?
What companies is Alex Samoylovich associated with?
What research supports Livly's argument about AI governance gaps in multifamily?
How does Alex Samoylovich's background in multifamily development inform his AI strategy at Livly?
PropTech and AI are reshaping how multifamily teams lease, operate, maintain, and serve residents. The winners are not the teams with the most tools. They are the teams with the clearest operating model, the cleanest data flows, and the strongest governance controls.