Decision Intelligence Architecture & Methodology
SemioTrace operates on deterministic principles. We believe that professional residential acquisitions cannot rely on black-box AI ratings or conflated data fields. Below is our formal methodology governing data verification, three-layer scoring, and financial modeling.
1. Property Identity & Entity Resolution
Every candidate is resolved to a single canonical Parcel ID (APN) and FIPS county code. When court foreclosure filings or probate records feature slight address variations (e.g. “Bayshore Vista Dr” vs. “Bayshore Vista Drive”), our normalization layer matches the legal description on file with the county recorder of deeds rather than relying solely on string comparison.
2. Strict Separation of Value Fields
To prevent fatal underwriting errors, SemioTrace strictly enforces field-level separation:
3. The Three-Layer Scoring Model
We reject single blended scores that obscure risk. SemioTrace computes three orthogonal dimensions:
Measures pure economic attractiveness: closed comp price spreads, distress urgency (pre-foreclosure / auction deadlines), days on market negotiation leverage, and favorable debt structure.
Evaluates alignment with the acquisition firm's explicit strategy: target MSAs, asset types, capital ceiling, and rehab budget limits. Hard deal disqualifiers are treated separately from scoring criteria.
Quantifies the recency, authority, and internal consistency of the supporting records. A property can possess high Opportunity (88) and moderate Data Confidence (65)—signaling analysts to verify title rather than discard equity.
4. Deterministic Economics Engine
Financial projections are calculated via audited algebraic formulas, never generated independently by LLMs. Our model projects All-In Basis, Net Profit, Margin on Cost, and Cash-on-Cash ROI across Downside, Base, and Upside scenarios with holding period cost reserves and debt service modeling.
5. Evidence States & Known Unknowns
Every material claim carries an explicit state: Verified, Supported, Conflicting, or Incomplete. Where vital information is unverified (such as interior structural conditions or junior municipal liens), SemioTrace explicitly highlights them as Known Unknowns rather than defaulting them to zero or masking them behind AI confidence.