The Three-Layer Underwriting Stack — Screening, Deep Analysis, Recommendation

Underwriting Is a Three-Layer Stack

Real estate deal underwriting is not a single activity. It's a three-layer stack — screening, deep underwriting, and recommendation — each with its own purpose, its own timeline, and its own analytical depth. Conflating the layers is one of the most common mistakes operators make. You don't deep-underwrite every deal that crosses your desk (you'd never get anything done); you don't screen-only every deal (you'd miss things that need depth). The discipline is knowing which layer applies to which deal.

Layer 1 — Screening

Screening is fast. The question is gating: does this opportunity even meet the operator's basic investment criteria? Standard screen elements: deal size band (does it fit our check size); asset class (does it fit our mandate); location (does it fit our geographic focus); headline returns (does the math even support our hurdle rate); deal-specific gates (institutional vs sponsor profile; off-market sourced vs broker-marketed; preferred sponsor vs unknown). Screening should take minutes to hours, not days. Most deals should fail screening — that's not a problem, that's the point. Operators who lack a disciplined screen end up deep-underwriting deals that were never going to fit.

Layer 2 — Deep Underwriting

Deals that pass screening graduate to deep underwriting. This is where the IC-memo-ready analytical package gets built: comparative market analysis (sales comparables plus lease comparables plus cap rate comparables for the submarket); capital stack stress testing (does the proposed stack work at higher debt cost, in less-favorable refinance environments); scenario modeling (multiple exit paths, multiple market trajectories, downside cases); operator track record review where applicable; legal and regulatory environment scan. Deep underwriting takes weeks rather than months when the work is AI-augmented — and the time compression is what makes it possible to deep-underwrite enough deals to actually find the good ones.

Layer 3 — Recommendation

The recommendation layer is where the underwriter has to make a call. Defensible recommendations come in three forms: YES (proceed with this deal as proposed), NO (decline this deal), or CONDITIONAL (proceed if certain assumptions hold; decline if they don't). The CONDITIONAL recommendation is often the most useful — YES if rent growth holds at 3.5 percent through stabilization; NO if it drops below 2.5 percent; build a re-underwrite at year 2 to validate. The recommendation traces back to the analytical backing in layer 2 and is structured to be defensible in front of an investment committee that may push back on every assumption.

Stress Testing — The Defensibility Question

Underwriting that only models the base case is not defensible. Institutional underwriting always layers downside scenarios — rent growth 200 bps below base; exit cap rate plus 100 bps; debt cost plus 150 bps; vacancy plus 500 bps from underwritten; construction cost plus 10-15 percent if applicable; schedule slippage plus 6 months if applicable. Run each downside individually AND in combinations. The interesting question is not whether the base case works — it's which downside scenarios break the thesis. Defensible underwriting always shows the breaking-point sensitivities, not just the upside ones.

What AI Augmentation Actually Does

AI augmentation doesn't replace the underwriter. It removes the friction at each layer. Screening: AI ingests deal collateral and surfaces the gating issues in minutes. Deep underwriting: AI pulls comparable transactions, runs the capital-stack modeling, generates the sensitivity tables, drafts the IC-memo sections. Recommendation: AI does NOT make the recommendation — that's human work — but AI assembles the analytical package the human relies on. The pattern that took a 10-person operation to a 3-person operation in a regulated-industry consulting engagement applies directly here.

Where the AZ Real Estate Licensee Scope Fits

Underwriting deliverables describe deal structures, capital stack architecture, lease provisions, contract terms, and transaction processes at the brokerage-education level appropriate for an experienced AZ Real Estate Salesperson with Wharton property-analysis training. That scope is legitimate — AZ Salesperson licensure authorizes preparing purchase agreements and lease agreements under broker oversight, negotiating on behalf of represented parties, providing market commentary, and discussing transaction-process options. Where legal advice is required (specific contract interpretation, securities law, tax structuring), the underwriter coordinates with the client's licensed attorneys.

Where the Series 82 Boundary Sits

Series 82 (private securities offerings representative) is targeted approximately 2027 and is NOT currently held. Until Series 82 is in place, underwriting deliverables are analytical work product for the client's own use in their investment decision-making — not investment advice, not solicitation of investors, not broker-dealer activity. Where the deliverable touches a specific securities transaction (a 506(c) offering memo, for example), the underwriter coordinates with the client's licensed securities attorneys and registered broker-dealer; the client's licensed counsel handles the securities-law-regulated activity.

Why Three Layers Matter

The discipline of separating screening, deep underwriting, and recommendation is what makes institutional underwriting scalable. Screening lets the operator process volume; deep underwriting lets the operator process complexity; recommendation lets the operator process defensibly. Combine them and the operator can sustain a deal flow that supports both the volume to find good deals AND the depth to make them defensible at IC.