LD Salmanson, CEO, Cherre

Institutional real estate data management remains broken not because firms lack analytics tools, but because no single source of truth connects their internal financial systems, operational data, and third-party market feeds. Cherre, founded by LD Salmanson in 2018, has built a data management and knowledge graph platform that addresses exactly that problem - used by large asset managers, banks, and insurance companies managing complex, multi-asset-class portfolios.

Key Takeaways

• The data problem precedes the AI problem. Most CRE firms cannot reliably answer what they own, where they own it, or how it is performing - not because they lack analytical tools, but because their data sits in disconnected systems with inconsistent definitions.
• Manual workarounds are expensive and fragile. Teams built around document collection, data reconciliation, and dashboard maintenance erode whenever staff turn over. Firms rebuild these processes every few years at significant cost.
• AI amplifies data quality, it does not fix it. Salmanson is explicit: institutional investors cannot rely on probabilistic AI outputs for compliance or public reporting. A reliable data layer must come first; AI functions downstream of it.
• Complexity determines necessity. A firm with two assets does not need Cherre. A firm managing multiple asset classes across geographies and reporting to public markets has no viable alternative to integrated data infrastructure.
• The knowledge graph changes the class of question. Cherre models relationships between assets, owners, markets, and transactions - not just static tables. Users can traverse connections, identify portfolio exposures, and surface risks that disconnected systems would never reveal.
• The 2026 roadmap is fully agentic. Salmanson's near-term development priority is a platform where users can build, call, and manage AI agents on the fly - all within SOC 1 and SOC 2 compliant infrastructure.

Working with CRE sponsors and institutional operators across more than $1.5 billion in transactions over 30 years, I have watched firms invest heavily in analytics while still reconciling basic performance questions manually. The conversation with Salmanson cuts directly to why that gap persists - and what resolving it actually requires.

The Real Problem Is Not Analytics - It Is Data Fragmentation

Commercial real estate has accumulated an impressive array of analytics platforms. What it has not built is a reliable foundation on which those platforms can operate.

Salmanson frames the issue directly: firms running ERP platforms, CRMs, underwriting models, and third-party data feeds are working with systems that use different definitions, different formats, and different update cycles. The result is not simply inefficiency. It is epistemic uncertainty - the inability to answer basic questions about performance, risk, and exposure with any confidence.

The traditional solution has been manual assembly. Teams collect documents, reconcile discrepancies, build dashboards, and maintain those dashboards until the people who built them leave. As Salmanson describes it, the process is "very expensive, very unreliable, very slow, very brittle." Many firms simply stop measuring altogether, reasoning - with a candor he clearly finds both amusing and alarming - that market appreciation is skill and market decline is circumstance.

What a Unified Data Layer Actually Does

Cherre's architecture centers on a knowledge graph - a structure that models relationships between entities rather than storing data in isolated tables. The graph covers assets, owners, markets, transactions, and financial performance across billions of entities in the built environment. Clients connect internal systems (Yardi, MRI, RealPage, Argus, JD Edwards, Salesforce) alongside third-party market data from providers including Green Street, RCA, Trepp, and Moody's - all feeding into a single, validated environment.

The practical result is a different class of question becomes answerable. Not just how is this asset performing against budget, but how does it perform relative to the market, what does comparable transaction pricing look like from RCA, and what does current debt look like from Trepp - queried simultaneously against a live system rather than assembled piecemeal by an analyst over several days.

Salmanson is careful about where AI fits in this architecture. Institutional investors reporting to public markets cannot rely on language model outputs without auditability. "I can't just say the LLM said it was X," he notes. Once the data layer is stable, AI earns a genuine role: automating variance analysis, generating commentary, identifying patterns across large datasets, and enabling natural language queries against structured data. But it operates on top of the infrastructure - not as a substitute for it. 

Why Adoption Is Concentrated Among Large Operators

Cherre's client base - large asset managers, banks, insurance companies, and technology firms building on the platform - reflects a structural reality rather than a pricing decision. Salmanson puts it plainly: a firm with two assets can walk them. A firm running multi-asset-class, multi-geography portfolios with public reporting obligations faces exponential complexity, and the cost of not having integrated data rises accordingly.

Deployment complexity varies sharply by client. A firm with well-documented ERPs and clean third-party data licenses can be onboarded in a day. A firm that has acquired sixteen platforms across a decade of growth, never standardized its chart of accounts, and operates three separate Yardi instances faces months of work before the infrastructure can function as designed. The platform accommodates both, but the timeline is a direct function of the mess underneath. 

What Changes Operationally Once Data Is Unified

Salmanson describes a platform moving toward fully agentic workflows - multiple AI agents running continuously against the unified data layer, surfacing alerts around lease expirations, flagging performance against underwriting, and suggesting actions based on user role and historical behavior. The demo shown during this episode illustrates how a portfolio manager can query holdings by market, asset class, or fund structure through a natural language interface - with the system interpreting organizational terminology automatically rather than requiring precise query syntax.

The 2026 roadmap extends this further: a fully agentic environment where users can build and call custom agents on the fly, all within SOC 1 and SOC 2 compliant infrastructure. The framing is not a dashboard company adding AI features. It is infrastructure on which CRE professionals can run customized analytical workflows without writing code - or waiting for an analyst to reconcile the inputs first. 

Frequently Asked Questions

What is a knowledge graph in commercial real estate?

A knowledge graph is a data structure that models relationships between entities - assets, owners, markets, transactions, and financial performance - rather than storing information in disconnected tables. In CRE, this means a query about an asset can automatically surface related market data, comparable transactions, and fund-level context without manual assembly. Cherre's knowledge graph covers billions of entities across the built environment and forms the foundation on which client-specific internal data is layered.

Why can't AI solve the CRE data problem directly?

AI models amplify whatever data quality they operate on. For institutional investors with compliance obligations and public market reporting requirements, probabilistic outputs from language models are not auditable - you cannot attribute a financial figure to "the model said so." AI becomes genuinely useful in this context once the underlying data is validated, standardized, and traceable from source to output. At that point it can automate analysis and enable natural language queries, but it cannot substitute for the data infrastructure itself.

How long does it take to implement a data management platform like Cherre?

Implementation time depends almost entirely on the client's existing data environment. A firm with well-documented ERP systems and clean third-party data licenses can be operational within a day. A firm that has grown through acquisitions without standardizing systems - running multiple ERPs with custom configurations and no unified chart of accounts - may require three to four months of deployment work. The platform is designed to handle both scenarios, but the timeline reflects the complexity of what already exists, not the platform itself.

Is this platform only relevant for very large institutional investors?

Cherre's core clients are asset managers above approximately $500 million AUM, along with banks and insurance companies. The reason is structural: complexity drives necessity. A small operator with a handful of assets can manage with simpler tools. Multi-asset-class, multi-geography portfolios with sophisticated reporting requirements reach a threshold where fragmented data is not merely inconvenient - it actively constrains decision-making. Salmanson notes that technology firms also represent a significant client group, building their own products on Cherre's platform rather than constructing the underlying data infrastructure themselves.

Learn More

If the data infrastructure question resonates with where your firm currently operates, explore how AI is reshaping the full CRE deal lifecycle - from deal sourcing through asset management and exit - at [LINK: GowerCrowd AI for commercial real estate]. For CRE sponsors building investor acquisition and capital formation systems, join the waitlist for the next AI Accelerator Executive Program.

Adam Gower, Ph.D. is the founder of GowerCrowd and is recognized as the most experienced practitioners in commercial real estate capital formation today. With more than 30 years and $1.5 billion in transactional experience - including serving as President of a Universal Studios development division overseeing $400 million in projects across Asia Pacific - Adam now helps CRE sponsors build AI-powered systems for investor acquisition, deal management, and capital deployment. His clients collectively manage over $45 billion in assets under management.