I model large real estate finance and construction projects — cost, the capital stack, returns, sensitivity — and I use agentic AI to do it faster. The AI accelerates the work; twenty-five years in the field is what verifies it.
I'm not selling you a framework. I'm a builder and a real estate practitioner who puts agentic AI to work on the projects I already understand.
Five disciplined methods I've used since college and in the field, now run AI-augmented and cross-checked. Estimates are approximate — not a bid, contract, or bindable price.
Read the method →Construction cost, the capital stack, IRR / equity multiple / cash-on-cash, and sensitivity analysis — on deals scaling into the hundreds of millions and billions. Investment-committee-grade, delivered in weeks.
AI handles the routine layer — research, data pulls, first-pass models — with a human in command. I review and verify every output against real field experience before it counts. AI still hallucinates; the trick is knowing what to check.
AI Twins as your interface to the world, the UPL / SEC / business boundaries, and keeper-in-the-loop protocols that hold the line.
Read →As the world automates, audit where your AI persona stands against your own compliance lines — and turn it into an AI Compliant Twin.
Read →The unfolding rulebook for AI Twins and AI Avatars — EU AI Act, NIST and ISO, and AI in law, business, and tax — and why public rules help each of us.
Read →A plain-language 2026 map of how the term AI Twin is used — 100 uses, with the regulated (law, medicine, finance) ones flagged. Educational, not a list of services.
Read →Titles, allusion & fair use — why a famous title is fair to allude to, tied to the compliant AI persona. General education, not legal advice.
Read →Narrow scope as a guardrail — how a small, no-advice thin agent becomes the friendly front door to an AI Compliant Twin.
Read →Goal-directed AI that acts across steps, with a human in command — in plain English.
Read →The DAIN Studios method for putting agents to work responsibly (Kruhse-Lehtonen & Hofmann).
Read →Reviewing a workflow for failure points, human gates, and compliance lines.
Read →Why a deliberately narrow-scope agent — one small job, no advice — is often the safer, more reliable design.
Read →