What is a Claude MHC park evaluation playbook? A Claude MHC park evaluation playbook is a structured set of prompts, file uploads, and Claude Project workflows that an operator uses to run due diligence on a manufactured housing community acquisition, covering rent roll integrity, trailing twelve month financial analysis, utility infrastructure liability, infill economics, and submarket comparable analysis. This guide gives MHC operators an actual playbook to copy, not a generic AI checklist. For the broader strategic context, see our pillar on AI manufactured housing.
Key Takeaways
- A Claude MHC playbook lives inside a Claude Project so every prompt has access to your underwriting standards, rent roll templates, and prior diligence files.
- The 1 million token context window in Claude Opus 4.7 lets you load full rent rolls, T12s, utility bills, leases, and a Phase I report in one workspace.
- Operators who run a structured Claude playbook on every park report 6 to 10 hours of analyst time saved per deal versus manual spreadsheet work.
- The biggest miss on MHC diligence is private utility liability, and Claude is well suited to flag it from utility bills, surveys, and seller representations.
- Use Claude for first-pass analysis only, then have a human operator validate every infrastructure and lot-count number before pricing the deal.
Why MHC Park Diligence Needs a Claude-Specific Playbook
Manufactured housing community diligence has a unique data problem. Sellers hand over rent rolls in spreadsheets that were last touched by a part-time bookkeeper, T12 statements that mix true operating expenses with owner draws, and utility bills that hide the real story on private water wells, septic fields, and sub-metering. Generic AI checklists tell you what to check. They do not tell you how to actually get the work done.
That is the gap a Claude playbook fills. Instead of a list of questions, you build a Claude Project that holds your underwriting model, your buy-box criteria, and the seller documents in one workspace. Then you run the same set of prompts on every park, every time. The playbook below is the structure operators are using in 2026 to take diligence from a 40 hour analyst slog to a 6 to 10 hour reviewed output.
Step 1: Set Up a Claude Project for the Park
Start by creating a Claude Project named after the park. Inside the Project, upload these files:
- Rent roll in Excel or CSV format, ideally the latest month plus the prior month for change tracking.
- Trailing twelve month operating statement with a column-level breakdown of revenue and expenses.
- Utility bills for the last 12 months, especially water, sewer, and electric.
- Survey or plat map showing lot count, lot lines, easements, and infrastructure footprints.
- Phase I environmental if available, plus any Phase II or remediation correspondence.
- Seller representations and the offering memorandum.
Inside the Project instructions, paste your buy-box: minimum lot count, maximum private utility exposure, target unleveraged yield, market rent ceilings, and your pricing methodology. This means every prompt downstream is grounded in your standards, not generic park advice. Operators who run this Project pattern at scale often pair it with the workflow described in our guide to automate rent roll with Claude Projects.
Step 2: Rent Roll Integrity Pass
The first prompt is the most important. Sellers regularly send rent rolls that overstate occupancy by counting park-owned-home rent as lot rent, or by leaving deceased and skipped tenants on the active list. Run this prompt:
"Read the rent roll. Identify every line that is not a true tenant-owned-home lot rent payer. Flag park-owned homes, RV pads, vacant lots presented as occupied, tenants more than 60 days delinquent, and any lots where rent is below 70 percent of the average. Output a clean count of true POH, true TOH, and vacant. Compare against the lot count from the survey and explain any discrepancy."
Claude will produce a reconciled rent roll. Cross check the lot count against the survey number. Any gap between the rent roll lot count and the survey lot count is a pricing issue. Report it.
Step 3: T12 Reconstruction
The trailing twelve month statement from the seller usually understates expenses on purpose. The next prompt strips it down:
"Compare the T12 to MHC operating expense benchmarks of $1,800 to $2,400 per occupied lot per year on a city water and sewer park, or $2,400 to $3,200 per occupied lot on a private utility park. Flag every line that is below the benchmark. Recommend which line items to underwrite higher. Identify any line that looks like an owner draw or capital expenditure mislabeled as operating expense."
Claude will produce a reconstructed T12. Use this as the input to your underwriting model, not the seller version. CRE investors looking for hands-on AI implementation support can reach out to Avi Hacker, J.D. at The AI Consulting Network.
Step 4: Utility Infrastructure Risk Pass
This is where MHC deals win or lose. Private water wells, lagoon septic systems, and master-metered electric on a park are the largest hidden liabilities in the asset class. Run this prompt:
"Read the utility bills, the survey, the Phase I, and the seller representations. Identify whether the park is on city water and sewer, well water, septic, or some combination. Estimate the cost to convert to municipal utilities if applicable. Flag any sign of system age over 30 years, undersized infrastructure for current lot count, or environmental liability from septic systems near surface water. Output a dollar range for utility-related capital expenditures over the next five years."
Claude will produce a utility risk memo. Read it carefully. The single biggest underwriting miss in MHC is treating a private utility park like a city utility park and ignoring the long-term capex bill.
Step 5: Submarket and Comparable Sales Analysis
Run a market pass to validate the in-place rents and the exit cap rate assumption:
"Search public sources for closed sales of MHC parks within 50 miles in the last 24 months. Output sale price, lot count, price per lot, and where available the cap rate. Compare the subject park to the comp set on lot count, utility setup, and condition. Output an estimated value range and call out the comp that is most analogous."
Claude can pull from its training data and any web access you have enabled. Validate every comp it produces against a real source like CoStar or industry databases before relying on it. According to CBRE Research, MHC remains one of the most resilient CRE asset classes in 2026, so comp scarcity is real and human validation matters.
Step 6: Infill and Revenue Upside Analysis
Most MHC value-add stories live in infill, which means filling vacant lots with new homes. Run this prompt:
"Given the rent roll and lot count, calculate the vacant lot count. Estimate the cost per home to infill at $35,000 to $55,000 depending on home spec, the time to fill at 1 to 2 homes per month, and the incremental NOI at the in-place market rent. Output an infill plan with capital required, timing, and incremental NOI by year. Stress test the plan with infill occurring at 50 percent of the assumed pace."
Claude will produce an infill underwriting case you can drop into your model. Use it as a check on the seller's pro forma upside story.
Step 7: Final Diligence Memo
The closing prompt synthesizes everything:
"Produce a 1 page diligence memo for the investment committee. Include: park summary, true lot count, reconstructed NOI, top three risks (with dollar exposure), recommended price range, and three deal breakers that would make us walk. Format as a memo with bullet points, not paragraphs."
This memo goes to the IC. The full Claude transcript and supporting prompts go in the deal file as backup. If you are ready to systematize this for a portfolio of acquisitions, The AI Consulting Network specializes in exactly this build-out. For deeper context on how Claude compares to other tools on lease and document analysis, see our comparison of Claude vs ChatGPT property valuation.
Common Pitfalls to Avoid
- Trusting the rent roll without reconciling to the survey lot count. Always force Claude to compare the two.
- Treating Claude output as final. Every infrastructure and lot count number gets human validated before pricing.
- Skipping the Project step. Without your buy-box loaded, Claude gives you generic advice instead of investment-grade analysis.
- Forgetting to load the Phase I. Septic and well environmental issues are deal killers and Claude needs the document to flag them.
Frequently Asked Questions
Q: Which Claude model should I use for MHC diligence?
A: Use Claude Opus 4.7 for the rent roll, T12, and utility analysis where reasoning depth matters. Use Sonnet 4.6 for the comp pulls and the final memo where speed and cost matter more. The 1 million token context window on Opus 4.7 means you can load the entire diligence package in one workspace.
Q: How long does this playbook take to run on a single park?
A: A trained operator running this playbook end-to-end on a park with a clean document set takes 6 to 10 hours, including human validation. The same diligence done manually with spreadsheets takes 30 to 40 hours.
Q: Can Claude actually catch private utility liability?
A: Yes, when you give it the utility bills, the Phase I, and the survey. Claude will not catch a liability that is hidden from the documents. Always supplement with a physical site visit and a third-party engineer for any park with private utilities.
Q: Does Claude replace the third-party diligence team?
A: No. Claude replaces the analyst layer that builds the underwriting case from the documents. You still need a real engineer for infrastructure, a real environmental firm for Phase I and II work, and a real attorney for title and closing. Claude makes the analyst layer 4 to 5 times faster.