Shipped This Week: Multi-Agent IC Pipeline
Five specialist AI agents (Underwriting, Market, Risk, Operations, IC) now collaborate to produce institutional-grade deal analysis in under 90 seconds. Live demo available.
This week we shipped the multi-agent pipeline that powers the Deal Analyzer.
The architecture is intentionally boring: five specialists, four running in parallel, one (IC) running sequentially as a reducer that consumes the others. Each agent has a strict JSON contract — the model is constrained to a typed output schema, and missing inputs return null rather than fabricated values. When the API key isn't configured, the orchestrator falls back to typed stubs so the rest of the platform keeps working.
What we wanted to avoid: an agentic system that fails opaquely, hallucinates specific facts, or produces inconsistent output formats. The cure is contract discipline — every agent is a function from inputs to a known JSON shape, and the orchestrator only trusts what the schema validates.
What this enables: an analyst can run a 48-unit multifamily deal end-to-end in 90 seconds and get back a draft IC memo with structured risk surfacing, market briefing, and a defensible recommendation. The analyst still owns the final judgment — the AI just produces the scaffolding faster than a human could.
Share kit
Post this to LinkedIn
Tip: copy the post text first, then click "Open LinkedIn share dialog" — paste into the dialog that opens. LinkedIn auto-renders the URL preview.
More Field Notes
Phoenix Multifamily — The Supply Story Is Decelerating
Phoenix multifamily deliveries peaked at 24K units in 2024. 2026 projects ~13K. The supply pressure that broke rents in 2024-2025 is materially abating — but the price discovery hasn't caught up yet.
Why We Build Tools That Don't Replace Judgment
A short note on the design philosophy behind LargeKite — AI augments analyst capacity rather than replacing committee judgment, and why that distinction matters more than the productivity gain.
The 5.85%-Cap-That's-Really-4.95% Pattern
A short methodology note on the systematic gap between broker pro forma cap rates and recast cap rates on multifamily deals — and what to do about it.