Offering Lens turns authorized DST PPMs and offering documents into structured, source-cited research—so your marketplace can publish richer listings, objective comparisons, and auditable answers without replacing the platform you already operate.
Most DST pages repeat a small set of sponsor-provided fields, then ask the visitor to interpret a long PPM alone. Offering Lens adds a structured intelligence layer between the documents and the experience you publish.
The AI locates and interprets the document set. Deterministic code performs arithmetic. Reconciliation catches conflicts. Human approval controls publication.
Map document versions, extract material facts, preserve page citations, and keep amendments visible.
Evidence before prose →Normalize stable fields while preserving the bases that make LTV, fees, and projections meaningfully different.
No hidden equivalence →Store formulas, inputs, units, and results for calculated leverage, fee allocation, and concentration metrics.
Arithmetic you can audit →Route conflicts, missing critical fields, ambiguous language, and stale sources into a focused human queue.
Uncertainty stays visible →Separate draft analysis from immutable investor snapshots. Publish only fields with the right approval status.
Your process, your gate →Serve the same canonical opportunity record to cards, detail pages, comparison tables, and internal tools.
One record, many views →A deliberate analysis sequence keeps the output complete, reproducible, and ready for review—not merely plausible.
Classify the PPM, amendments, loan materials, appraisals, rent rolls, and supporting documents.
Version-aware mapSeparate agents analyze structure, fees, debt, property operations, conflicts, and liquidity against one evidence model.
Narrow expert scopesDetect conflicts, superseded values, missing citations, calculation mismatches, and material review items.
No silent resolutionCreate a canonical record, approve the right fields, and deliver an immutable research snapshot to your listing surface.
Human publication gateOffering Lens is designed around a small, stable opportunity record that your existing frontend can consume. Start with one offering detail page, then expand to comparisons, research Q&A, and internal review.
{
"display_name": "Example DST Opportunity",
"analysis_status": "approved_investor",
"key_metrics": [
{
"field_id": "debt.loan_to_value.stated",
"value": 55.0,
"basis": "appraised_value",
"value_type": "disclosed",
"source_fact_id": "fact_ltv_1"
}
]
}The product boundary is part of the architecture. Offering Lens is designed to surface evidence and uncertainty—not make a suitability decision, guarantee exchange eligibility, or auto-publish unreviewed output.
Material disclosed facts require at least one supporting source with a 1-indexed PDF page and a concise excerpt.
Disclosed, calculated, interpreted, and not-found fields cannot silently collapse into one undifferentiated answer.
Investor views use only approved fields from an immutable snapshot—not the live draft analysis state.
A practical starting point for marketplace operators evaluating AI-powered DST document analysis and listing research.
It is a structured research layer between authorized offering documents and the listing experience. It turns document evidence into normalized facts, reproducible calculations, review items, comparisons, and approved investor-facing content.
That is the intended architecture. A normalized opportunity record can feed existing cards, detail pages, comparison tools, and research rooms while your platform retains its own brand, access controls, and document policies.
The workflow first maps documents and versions, then runs focused analysis across offering structure, fees, debt, property operations, management conflicts, and liquidity. A reconciliation stage checks conflicts and missing support before a canonical record is created.
No. Offering Lens is designed to describe and compare disclosed facts, assumptions, calculations, and approved review context. It does not rank a winner or make a personalized recommendation, suitability conclusion, or best-interest determination.
No. The system records how the offering represents its intended tax treatment, preserves the operative source language, separates disclosure from interpretation, and flags material ambiguity for appropriate human review.
Candidate values are preserved, explicit supersession is linked when supported, and unresolved conflicts remain visible in the review queue. The system is designed not to silently choose the more convenient value.
Yes. Because each opportunity uses stable field IDs and linked evidence, the same approved record can support objective comparison and source-cited answers. Bases and dates remain visible so similar-looking figures are not treated as equivalent without support.
Bring one authorized offering package. See how it becomes a source-cited opportunity record, review queue, comparison column, and investor research page.