Treating AI like a junior: building trust in AI

"AI is very exciting to everybody... it's opening up a lot of opportunities and it's somewhat leveling the playing field."

That's Adam Siegel, Vice President of Product Growth at Crexi, talking about what he is seeing across commercial real estate.

Adam has been at Crexi since August 2024, where he works on growing the company's intelligence platform by bringing better data onto the platform, acquiring relevant datasets and building workflows that help users get things done. Before Crexi, he spent more than 11 years at CBRE, managing the day-to-day operations of two of the country's largest capital markets teams, responsible for more than 250 deals a year and approximately $2–3 billion in annual consideration.

Adam and Jonas get into what AI is actually changing for brokers, investors, researchers and other commercial real estate professionals — from faster market research and more granular data to AI-powered workflows, MCP servers and the risks of trusting AI with high-stakes decisions.

In this episode

AI is leveling the playing field. AI is giving commercial real estate professionals faster access to data and allowing them to work at a velocity that was previously difficult to achieve. Adam talks about brokers, appraisers, lenders and private investors using AI to look at more deals, understand markets, check buy boxes and research things like market rents and new construction without spending hours digging through reports, emails and websites.

Breaking down the data silos. Adam sees a major opportunity in connecting data that currently lives in different places. The goal isn't just to ask an AI a question and get an answer, but to let professionals interact with their data, drill down into it and keep asking better questions. He compares it to the evolution from canned reports to working directly with raw data in Excel and pivot tables.

MCP: scam or the next step? Jonas brings up an industry contact who called MCP a "scam." Adam pushes back on the idea, arguing that MCP is further along the AI journey and becomes valuable for people who want to manipulate data, access it directly and keep drilling down rather than settle for a simple question-and-answer experience.

The accuracy problem in high-stakes real estate. AI can move incredibly quickly, but Adam isn't ready to trust it completely with underwriting. He talks about the possibility of AI missing a critical clause, environmental issue or other detail in a transaction — potentially creating a multimillion-dollar mistake. The models are improving quickly, but the confidence level still needs to catch up.

From AI chat to useful workflows. Adam thinks the next stage of adoption is about building AI into people's daily workflows rather than simply giving them another chatbot. Crexi is working on tools that can function as a research assistant, combine proprietary and user-provided data and let users drill down to the street level.

Why commercial real estate needs its own AI. Adam argues that general-purpose models such as Claude don't have the same commercial real estate context and proprietary data that a specialised platform can bring. The opportunity is to tune AI agents specifically for the industry and give users insights they might not otherwise see.

One dataset, multiple outputs. Adam talks about Crexi's Create product and the idea of using the same underlying information across the lifecycle of a deal — from a BOV and due diligence through to an offering memorandum, website and marketing. The objective is to save small amounts of time repeatedly, while also allowing the system to learn how an individual user works.

AI and the future of multifamily. Adam believes multifamily is particularly well positioned to benefit from AI, including for smaller operators. He sees opportunities well beyond acquisitions and dispositions, including understanding rents, maintenance, property management and collections. He expects some asset classes to move faster than others, but believes every part of real estate will benefit from AI in some form.

Building with AI instead of just talking about it. Adam is using AI to turn product ideas into working, clickable prototypes that can incorporate live Crexi data from Snowflake. Rather than spending hours explaining an idea to developers and designers, he can hand them something tangible that communicates his vision and gives them a head start.

AI adoption at Crexi. Adam describes how Crexi's engineers, product managers, designers and leadership are increasingly using AI as a force multiplier. He says the company has taken a top-down approach to becoming AI-forward, with the ambition to be more than a marketplace and instead bring AI solutions to the industry.

The callout: Crexi's engineering team. Every episode of Net Effective ends with a guest recognising somebody doing great work with AI. Rather than naming one individual, Adam gives the nod to Crexi's engineers, product leaders and executives for embracing AI as a force multiplier and helping rebuild the company's technology stack around it.

Mentioned in this episode

Adam Siegel, Crexi · Dwellsy · CBRE · Claude · ChatGPT · MCP · Snowflake · JLL · Create · Crexi's research AI tools · BOVs · Offering Memorandums · Los Angeles multifamily · Multifamily · Commercial Real Estate

Net Effective is a conversation with the people running residential rentals about how they are actually using AI. New episodes weekly, about 30 minutes. Subscribe at neteffective.show.

Connect with Adam

Adam is available on LinkedIn under Adam Siegel at Crexi and says he's happy to connect, answer questions and talk about what Crexi is building.