Research
Field notes from the work behind the agent.
No generic AI forecasts. These notes cover workflow boundaries, evidence, review, and deployment decisions inside institutional real estate.
Latest notes
- Portfolio intelligenceStart with exception queues, not a chatbotA practical operating model for reconciling fund data, surfacing discrepancies, and routing controlled work.
- Operating modelWhere AI agents belong inside a real estate investment fundA workflow test for deciding what to automate, what to leave in the system of record, and who must own release.
- Fund operationsAn AI-assisted quarter-end close without a black boxHow to preserve transaction lineage, exception handling, and fund-accounting approval while reducing manual assembly.
- Fund operationsA distribution waterfall should explain itself tier by tierWhat an agent can prepare, what the LPA still governs, and why reviewable math matters more than a fast answer.
- Investor relationsThe investor answer is only correct inside the right boundaryFund, investor, document, and side-letter scope for AI-assisted LP reporting.
- Deal workDiligence is not done until the source survives the conclusionA practical evidence spine for VDR extraction, sponsor screening, conflicts, and committee materials.
- Portfolio intelligenceWhen the model and property system disagreeHow to produce a point-in-time portfolio answer without averaging away the conflict.
- Brokerage operationsWhat a brokerage voice agent must hand back to the brokerQualification, consent, transcript, routing, and the record required after an inbound call.
- Operating modelKeep the system of record. Automate the work above it.A safer deployment sequence for firms that cannot afford a replacement project disguised as an AI pilot.
Calculators
Planning tools only. Outputs are estimates, not investment, legal, tax, or accounting advice.