Use Claude Code in a commercial real estate business
I’m a solo commercial real estate developer, owner, and syndicator focused primarily on retail centers and industrial deals across multiple states (mostly Western US). Task: I’m a solo commercial real estate developer, owner, and syndicator focused primarily on retail centers and industrial deals across multiple states (mostly Western US). Typical deal sizes range from roughly $5 million to $30 million. Like a lot of people in CRE, I’ve relied heavily on Argus over the years. In practice, that usually meant sending things out to third-party analysts to run the models because Argus is time-consuming, expensive, and not exactly something I wanted to become a full-time expert in. Over the past year or so, I started experimenting more seriously with AI tools. First Claude and then CoWork because of its Excel integration, which has been extremely useful. More recently, I’ve been getting into Claude Code, and that has changed the game for me. I’ve now been able to build DCF models that are getting close to Argus-level flexibility and reliability, at least for the types of deals I’m underwriting. I started with a single-tenant industrial model, then built out a multi-tenant industrial model focused on small/mid-bay product. Now I’m working on adapting the structure for multi-tenant retail. The models are still Excel-based, but CC has helped me build far more dynamic logic around rent rolls, reimbursements, downtime, renewal probability, market leasing assumptions, tenant improvements, leasing commissions, debt, exit assumptions, and sensitivity outputs. The biggest difference is that I can now customize the model exactly around how I think about a deal instead of forcing everything through a rigid third-party process. And it takes minutes, not hours/days to do! I’m not saying Argus is dead, especially for institutional shops or highly standardized reporting. And certainly for very large deals, portfolio's, etc. But for a solo operator like me, I’m starting to think I may be done relying on Argus and paying outside analysts pretty soon. Curious if anyone else in CRE is going down this path. Are you using AI to build or audit underwriting models? Have you been able to replace parts of your Argus workflow? Or do you still think Argus remains necessary once deals get complex enough? I'm wondering what else I can use CC to greatly improve my efficiency, since time is my #1 constraint.; tools: Claude Code. Human role: The original post does not say how a person stayed involved.; limitations: Prerequisites include proficiency in Excel and commercial real estate underwriting concepts. Claude Code requires manual verification of generated formulas and logic, as AI may introduce subtle errors in complex financial models. Integration uncertainty exists regarding long-term model maintenance and version control. Human approval is mandatory for all underwriting assumptions and final model outputs. Source reliability is secondary, based on a single user anecdote without peer review..
Tools
Published on Agent Work Atlas:
Overview
I’m a solo commercial real estate developer, owner, and syndicator focused primarily on retail centers and industrial deals across multiple states (mostly Western US).
Workflow
I’m a solo commercial real estate developer, owner, and syndicator focused primarily on retail centers and industrial deals across multiple states (mostly Western US). Typical deal sizes range from roughly $5 million to $30 million. Like a lot of people in CRE, I’ve relied heavily on Argus over the years. In practice, that usually meant sending things out to third-party analysts to run the models because Argus is time-consuming, expensive, and not exactly something I wanted to become a full-time expert in. Over the past year or so, I started experimenting more seriously with AI tools. First Claude and then CoWork because of its Excel integration, which has been extremely useful. More recently, I’ve been getting into Claude Code, and that has changed the game for me. I’ve now been able to build DCF models that are getting close to Argus-level flexibility and reliability, at least for the types of deals I’m underwriting. I started with a single-tenant industrial model, then built out a multi-tenant industrial model focused on small/mid-bay product. Now I’m working on adapting the structure for multi-tenant retail. The models are still Excel-based, but CC has helped me build far more dynamic logic around rent rolls, reimbursements, downtime, renewal probability, market leasing assumptions, tenant improvements, leasing commissions, debt, exit assumptions, and sensitivity outputs. The biggest difference is that I can now customize the model exactly around how I think about a deal instead of forcing everything through a rigid third-party process. And it takes minutes, not hours/days to do! I’m not saying Argus is dead, especially for institutional shops or highly standardized reporting. And certainly for very large deals, portfolio's, etc. But for a solo operator like me, I’m starting to think I may be done relying on Argus and paying outside analysts pretty soon. Curious if anyone else in CRE is going down this path. Are you using AI to build or audit underwriting models? Have you been able to replace parts of your Argus workflow? Or do you still think Argus remains necessary once deals get complex enough? I'm wondering what else I can use CC to greatly improve my efficiency, since time is my #1 constraint.
Reported result
Actual outcomes are not measured or reported in the source. The author claims models approach Argus-level flexibility and reliability for their specific deal types, but no quantitative metrics, error rates, or financial impacts are provided. Proposed acceptance check: Validate model outputs against a known historical deal or third-party audit to confirm calculation accuracy and logic integrity before relying on them for investment decisions.
Human role
The original post does not say how a person stayed involved.
Limitations
Prerequisites include proficiency in Excel and commercial real estate underwriting concepts. Claude Code requires manual verification of generated formulas and logic, as AI may introduce subtle errors in complex financial models. Integration uncertainty exists regarding long-term model maintenance and version control. Human approval is mandatory for all underwriting assumptions and final model outputs. Source reliability is secondary, based on a single user anecdote without peer review.
Try it
Act as a commercial real estate underwriting assistant. I am a solo developer focusing on $5M-$30M retail and industrial deals in the Western US. I need to build a dynamic DCF model in Excel using Claude Code to replace my reliance on Argus. Inputs: [replaceable inputs: rent roll CSV, lease terms, market assumptions, debt structure, exit cap rate]. Steps: 1. Analyze the provided rent roll and lease data. 2. Generate Python or VBA code to structure an Excel model with dynamic logic for rent, reimbursements, downtime, renewal probability, TI/LC, debt, and exit assumptions. 3. Create sensitivity tables for key variables. 4. Output the model structure and code for me to implement in Excel. Deliverables: A detailed model architecture document and code snippets for Excel integration. Quality Check: Ensure all formulas are transparent and auditable. Safeguards: Do not finalize any model. Run a dry test with a sample deal first. I will manually verify all calculations against a known benchmark before using the model for real underwriting. Ask clarifying questions if lease terms or market assumptions are ambiguous.
Source reports are not independently verified. Publication date is the catalog publication date, not the original article date.