Red-team student-housing deals before IC
A student-housing team used Claude to check underwriting assumptions, score deals across five areas, and draft a credit call before human review.
- What they did
- They built a multi-agent flow inside Claude Projects for an acquisition platform. One agent screened incoming deals, another pulled figures from the OM, T-12, and rent roll, and a later agent handled underwriting. A CCO agent then applied the same credit policy each time, pushed back on assumptions, scored five weighted dimensions, and could change the recommendation with a written explanation. Each deal lived in its own thread, and the outputs were passed along in structured packets before a person sent anything to the investment committee.
- How it went
- They said the CCO step was the biggest improvement because it made the credit review consistent. Final reports also synced to Google Drive for partners. Their reported turnaround dropped to under an hour for most deals, but the process still needed human approval and was not fully automated.
- Tips
- Keep the approval checkpoint before any submission or other irreversible step. Use a fixed scoring rubric so each deal is judged the same way, and keep one conversation per deal to avoid mixing context.
Try it yourself with Claude
You are the credit-policy reviewer for a student-housing acquisition team. I will paste the OM, T-12, rent roll, and our draft assumptions for [deal name]. First, check the assumptions for gaps or weak points, then score the deal from 1 to 10 across [five dimensions you choose, such as market, property, sponsor, financials, and risk]. Finish with a clear credit recommendation: approve, revise, or decline, and explain why. Before anything is sent to investment committee or uploaded anywhere, pause and ask for my approval.
Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.
Get 3 agent uses like this every morning
With the day's top AI news. Free, and one click to unsubscribe.
Related agent uses
Ask the support bot for a refund
They used Claude’s in-app help bot to seek a subscription refund and received approval by email immediately.
Build a daily brokerage research brief
A Claude assistant pulls a Schwab snapshot, checks news and Fed updates, and turns it into a morning report tied to the user’s holdings.
Check state unclaimed-property databases for old names
Have an agent look for forgotten funds across official state databases using your current and past names.
Compare hospital codes to Medicare before you negotiate
Claude helped price a hospital bill by Medicare rates so the patient could challenge the charges himself.
Turn bank statements into a spending dashboard
They use Claude to combine monthly statements, map spending, and review portfolio allocation in one place.
Use Claude to file tenant complaints from public records
Claude checked NYC building records, prepared 14 housing complaints, and submitted them for one tenant.