Pillar IV · Briefs · Forthcoming Q3 2026 · Free
What Claude can actually do on a CRE deal. And where it breaks.
A practitioner guide to using Claude in CRE underwriting work – rent roll review, lease abstraction, operating statement normalization, market comp summarization, and document Q&A. Includes tested prompt sequences, honest output benchmarks, and the failure modes that matter before you trust the output.
What this brief covers
~16pp
Six workflow modules, prompt sequences, output benchmarks, failure-mode taxonomy.
Free
Free with Dispatch. No payment required.
6 modules
Rent roll, lease abstraction, OS normalization, comp summarization, document Q&A, narrative drafting.
Who this is for
Analysts building underwriting packages
Replace two hours of rent-roll reformatting with a 10-minute Claude workflow. Know which output fields to verify manually and which are reliable enough to pass through.
Asset managers reviewing leases
Abstract 20 leases in the time it used to take to abstract three. The brief tells you which abstraction fields Claude gets right consistently and which require a second read.
Acquisitions professionals
Summarize a broker's comp package before the IC meeting. Normalize operating statements from seller-prepared formats into your firm's template. Draft the narrative sections of the investment memo first-pass.
Lenders and debt brokers
Run document Q&A against a loan agreement, appraisal, or environmental report before the borrower's counsel bills for the same exercise. Know what the tool can surface and what it will miss.
Outline
I. How Claude Handles CRE Documents
Context window mechanics, file upload behavior, what Claude preserves and loses across long documents, and the setup choices (Projects vs. single session vs. API) that affect output quality for CRE-specific tasks.
II. Rent Roll Review
Prompt sequence for ingesting a multi-tenant rent roll, extracting lease expiration schedules, flagging below-market leases, and producing a summary table. Output format tested. Known failure modes: merged cells, non-standard column headers, handwritten annotations.
III. Lease Abstraction
Field-by-field abstraction prompt structure. Which fields (base rent, term, option language, renewal notice windows, assignment provisions) extract reliably. Which fields (percentage rent triggers, operating expense exclusion definitions) require verification. Output template included.
IV. Operating Statement Normalization
Converting seller-prepared operating statements into a standard format. Identifying non-recurring items, management fee adjustments, below-the-line inclusions. Tested against trailing 12-month and calendar-year formats.
V. Comp Package Summarization and Document Q&A
Summarizing broker comp write-ups into usable market data. Running structured Q&A against loan agreements, environmental phase-one summaries, and appraisals. What Claude finds. What it doesn't.
VI. Narrative Drafting and Failure-Mode Taxonomy
Using Claude for first-pass investment memo narrative, deal summary sections, and lender-facing descriptions. Full failure-mode taxonomy: hallucinated figures, confident-but-wrong abstractions, context-window truncation, formatting collapse on complex tables.
FAQ
Do I need the paid Claude plan?
For the file-upload workflows and extended context needed for full lease abstraction and rent roll review, yes – Claude.ai Pro ($20/month) or an API key. The free tier has context limits that will truncate longer documents. The brief notes where free-tier use is sufficient and where it isn't.
Does this work on PDFs or just text?
Both. Claude.ai supports PDF uploads directly. The brief covers both paths – direct upload and text extraction first – and notes where PDF parsing introduces extraction errors on specific document types (e.g., scanned rent rolls, older appraisal formats).
What about data confidentiality?
Out of scope for the brief – that's a firm policy and legal question. The brief focuses on workflow and output quality, not data governance. Review Anthropic's usage policies and your firm's AI use guidelines before uploading deal-specific documents.
How is this different from the ChatGPT companion guide?
The two guides are parallel in structure but cover model-specific behavior. Claude and ChatGPT handle long documents, table extraction, and instruction-following differently. The briefs are designed to be read together or independently.
Will this be updated as models change?
Yes. Point updates will track major model releases and any meaningful changes to Claude.ai's file handling or Projects feature. Free for 12 months with Dispatch sign-up.
Is this an endorsement of Claude or Anthropic?
No. It's a practitioner assessment. Where Claude performs well on CRE tasks, the brief says so. Where it breaks, the brief documents the failure mode. Valore Registry has no commercial relationship with Anthropic.
Delivery
Free with Dispatch sign-up. PDF, ~16 pages, searchable. Free point-update releases for 12 months. Informational only – not legal, tax, financial, or investment advice. AI model behavior changes with each release; verify outputs against source documents before relying on them in any transaction.
Quarterly refresh. Free re-download for 12 months from purchase.
14-day refund if the file is materially different from what was described, corrupted, or not delivered correctly.
Or get this in All-Access Bundle for $997 — save ~44% vs à la carte.See all bundles →