Guest: Michael Episcope, co-CEO, Origin Investments

 

How Origin Investments Embedded AI Across the Firm

Michael Episcope of Origin Investments explains why AI adoption is now the biggest competitive risk in CRE.

 

Enterprise AI adoption in commercial real estate is no longer a technology question - it is an operations question. Origin Investments, the Chicago-based multifamily sponsor with nearly $2 billion in assets under management, has spent the past six months embedding an enterprise AI layer called AskSimon across every division of the firm, connecting structured data, Microsoft SharePoint, and a switching mechanism that routes queries to the best available LLM. The operational result: leadership can pull real-time portfolio data, cash flow statements, and fundraising figures without interrupting heads of divisions. The competitive implication: firms that delay are falling behind.

 

Key Takeaways

 

• Enterprise AI is already live at Origin. The AskSimon system went from implementation to mission-critical in under three months, integrating across Microsoft products with a switching mechanism that routes queries to Claude, GPT, Gemini, or other LLMs depending on task type.

• Data quality is the real gating factor. Michael Episcope rates his own divisions from A to F on data organization - and says firms that expect to layer AI over unstructured data without heavy lifting are kidding themselves.

• AI will not replace staff, but people who use it will replace those who don't. Origin has told its team directly: the risk is not the technology, it is the colleague who has learned it.

• Structured data is a prerequisite, not a feature. Episcope is explicit that AI systems require a single source of truth. A "sloppy lake of information" produces low-quality outputs regardless of which model sits on top.

• Investor portals may be the next interface to change. Episcope foresees conversational dashboards replacing static portals - investors asking natural-language questions about their positions rather than navigating reports.

• Sentiment data signals a CRE recovery. Two years ago, 8% of wealth managers wanted to increase real estate allocations. Today, that figure is 54%, with 70% of institutions planning to allocate more.

• The window for early-mover advantage is finite. Episcope puts it at roughly 18 months to 2 years before AI adoption becomes table stakes rather than differentiation.

 

Adam Gower has spent more than 30 years in commercial real estate, completing over $1.5 billion in transactions and serving as President of a Universal Studios development division overseeing $400 million in projects across Asia Pacific. His clients collectively manage over $45 billion in AUM. The question of how CRE operators restructure around AI - and what happens to those who don't - has been the central subject of his research and his work with sponsors over the past two years.

 

Why Origin Built AskSimon - and What It Actually Does

Most real estate firms experimenting with AI are doing the same thing: opening ChatGPT or Claude in a browser tab and asking questions. Origin Investments went a different direction.

 

The firm engaged AskSimon, a platform built specifically for small and medium-sized financial firms, which integrates directly into Origin's Microsoft suite. The system connects SharePoint, OneDrive, and other internal data stores, then routes queries through a switching mechanism that selects the most capable LLM for each request - image generation, deep research, and data retrieval each go to different models.

 

The business case became clear fast. As Episcope described it: "If I need to know how our portfolio's doing, I have to email Mark Turner, our Director of Investments, and then he has to go through in his file and get all this stuff, and then send me, and that's not a good use of his time. Or if I need something from legal or accounting, I had to do that. Now I just go into the system."

 

The system went from implementation to indispensable in roughly two to three months. Origin's leadership now queries fundraising figures month-to-date, cash flow statements, and portfolio performance without routing requests through division heads. Those division heads, freed from reporting overhead, focus on what they were hired to do.

 

Much like the productivity premium sponsors get from using AI, the security dimension matters as much as the functionality. Episcope was direct on the point: external LLMs like Claude or ChatGPT, used without an enterprise wrapper, feed data into systems that do not keep it confidential. AskSimon operates entirely within Origin's Microsoft environment, with no data shared externally.

 

Data Organization Is the Heavy Lifting Nobody Talks About

Origin has been six months into the implementation at the time of recording, with weekly technical updates and monthly divisional meetings focused on data cleanup. Episcope rates his own divisions on data quality - some are an A, some a B, some a C, and some an F. The variance matters, because AI can only interrogate data it can find and parse.

 

His framing was precise: "For any AI system to work, you have to have structured data underneath the system, and then you layer agents and things on top of that, and you have to point it to the right places where it's going to get the data from."

 

Firms expecting to deploy an AI layer over scattered files - documents on desktops, OneDrive accounts, and SharePoint in different organizational states - and receive coherent outputs will be disappointed. The data structuring work precedes any useful intelligence. There is no shortcut.

 

The implication for sponsors considering AI adoption is straightforward: the first project is not the AI system. As data infrastructure is vital for AI readiness, the first project is the data audit. Understanding where information lives, who owns it, and in what format it is stored determines what an enterprise AI system can actually do.

 

Origin's internal AI integration team - a dedicated group of three people planned by year-end - exists precisely to close this gap. They map divisional data quality, handle connectivity issues between data sources, and field queries from the rest of the organization about what the system can and cannot do.

 

The Human-in-the-Middle Requirement and What It Means for Hiring

Episcope drew a consistent distinction throughout the conversation between AI as a productivity tool and AI as a replacement for judgment. His position was that the two categories are not in tension - they are sequential. AI accelerates output; an experienced professional validates it.

 

The hallucination problem illustrated the stakes. Episcope described an incident researching Qualified Opportunity Zone legislation on open AI: ‘It gave me the exact opposite answer of what I wanted... I asked, “where did you find this?” And the location the bot responded with did not exist.’

 

For research-oriented outputs - investment memos, market analyses, regulatory summaries - someone senior enough to recognize a wrong answer must sit between the AI output and the final product. That is a structural constraint.

 

Origin's message to its staff has been correspondingly unambiguous. As Episcope put it: "AI will not replace you, but somebody who knows AI and knows your job will. And if you're not going to learn it, and you're not going to embrace it in here, then it's time to move on."

 

The new hire category this creates - Origin has already posted for an AI Integration Lead - did not exist four years ago. Firms building AI into operations are creating roles that require dual literacy: deep domain knowledge in whatever function they support, combined with the ability to interrogate, direct, and validate AI outputs. The GowerCrowd AI in Real Estate Accelerator executive program provides the training teams need.

 

How AI Is Changing the Investment Thesis for Multifamily

Origin's Multilytics platform, built starting around 2019-2020, ingests billions of data points monthly - population statistics, migration data, employment figures, interest rates, rental concessions, and comp set occupancy - to produce a rent forecast. Episcope describes it as the most accurate forecasting tool available at the one-to-three year range, while acknowledging that dispersion widens beyond that horizon.

 

The AI dimension intersects with Multilytics in a specific way: the platform is picking up real-time data that reflects present conditions, but macroeconomic shifts driven by AI-induced labor market changes have not yet fully appeared in historical datasets. Episcope's position is that Sunbelt migration, tax-friendly state preferences, and business relocation trends remain the primary drivers of multifamily demand - and those trends are durable regardless of AI's labor market effects.

 

His sector view on AI's property-level impact was concrete. He cited Elise AI as one example of a company attempting to automate property management functions. The financial logic is direct: converting a 4.5-cap asset to a 5-cap through expense reduction, without changing the purchase price, is significant value creation. 

 

On the macro picture, Episcope offered a framing he finds more useful than the dystopian-versus-utopian debate: "AI is like a Lipitor for the large, bloated tech companies, and it's steroids for the entrepreneur." The headline layoffs at large technology firms draw attention; the dispersed hiring at smaller firms enabled by AI does not. The job market statistics, he argues, reflect the latter more than the former.

 

The geographic thesis is unchanged. California's wealth tax proposals and persistent regulatory environment, combined with Illinois's fiscal pressures, continue driving corporate and individual migration toward Texas, Tennessee, and Florida. Those are the markets where AI-enabled job creation will concentrate, and therefore where multifamily demand will grow.

 

Frequently Asked Questions

How is Origin Investments using AI internally?

Origin has deployed an enterprise AI layer called AskSimon, integrated across its Microsoft suite including SharePoint and OneDrive. The system uses a switching mechanism to route queries to different LLMs depending on the task - Claude for certain research functions, GPT for others, Gemini for others still. Leadership uses it to pull real-time portfolio data, cash flow statements, fundraising figures, and market information without routing requests through division heads. Origin also maintains Multilytics, a proprietary machine learning rent forecast tool that ingests billions of data points monthly to model rental demand across its target markets.

 

What is the biggest obstacle to AI adoption for CRE sponsors?

Data organization. Episcope understands that AI systems require structured, accessible data to function. Firms with information scattered across desktops, personal cloud accounts, and SharePoint in inconsistent formats cannot expect coherent outputs from an AI layer. The preparatory work - auditing where data lives, standardizing formats, building proper data governance - precedes any useful AI deployment. Most firms underestimate this requirement. Origin has spent six months on it with weekly technical meetings and dedicated monthly divisional reviews, and still rates some divisions at a failing grade on data organization.

 

Will AI reduce headcount at CRE investment firms?

Episcope's position is that AI will not reduce headcount directly, but firms that adopt it effectively will be able to operate at greater scale with the same or smaller teams - and firms that do not adopt it will face competitive displacement from those that do. Origin has been explicit with its staff that the risk is not AI replacing them, but a colleague who has learned AI replacing them. The firm is hiring an AI Integration Lead - a role that did not exist four years ago - and is building a dedicated three-person AI integration team. The net effect on employment in CRE is likely a shift in skill requirements rather than a reduction in total positions.

 

How should CRE sponsors think about AI and investor communications?

Episcope sees investor-facing AI applications as the next frontier after internal operational efficiency. His near-term view is that traditional investor portals - static dashboards with quarterly report PDFs - may give way to conversational interfaces where investors ask natural-language questions about their holdings and receive direct answers. Origin has not yet deployed this, but the internal infrastructure being built with AskSimon creates the foundation for it. On the communication side, Episcope acknowledged that documenting and sharing how the firm uses AI - in the same way Origin educated investors about IRR limitations - could serve as a capital formation tool, distinguishing the firm from peers who have not yet engaged with the technology.

 

Go Deeper

If you are evaluating how to structure AI adoption at your firm, start with the operational layer before the investor-facing applications. The data audit comes first. Use the contact page on the website to let us know how we can help you.

To build the systems Episcope describes - enterprise AI integration, structured data pipelines, and AI-powered investor communications - inside your own organization, the GowerCrowd AI Accelerator Program provides a live implementation track for CRE sponsors. Learn more: AI in Real Estate Accelerator executive program.

 

Adam Gower, Ph.D. is the founder of GowerCrowd and one of the most experienced practitioners in commercial real estate capital formation. 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. gowercrowd.com