Collaborative Deal Research Workspace

Shipped

An AI-powered workspace where analysts research, extract, and analyze deal data — backed by a structured deal graph, indexed documents, spatial intelligence, and Slack integration.

Started
November 2025
Stack
TypeScript, Next.js, Firestore, BigQuery, Gemini, Google Cloud Run, Vercel, Mapbox, Slack API

The Problem

Deal analysis was fragmented: demographics in one place, competitive data in another, documents in email attachments, site data in a different system. Analysts spent more time finding information than analyzing it. The AI tools available couldn't produce reliable analysis because the underlying data was scattered and unstructured.

The special thing about this tool isn’t the AI. It’s the data architecture underneath that makes the AI finally useful.

Most AI tools fail for a simple reason: the data feeding them is scattered, inconsistent, and unstructured. You can’t bolt a model onto chaos and expect insight. At best you get pretty slop. At worst, hallucinated conviction.

Five design decisions drive everything:

Unified deal graph. Demographics, supply, competition, trade areas, foot traffic, financials, operating benchmarks — all normalized, all linked, all sourced. Most firms operate with PDFs, Excel files, and tribal knowledge. This turns that into structured intelligence.

Documents become data, not attachments. Appraisals, offering memoranda, feasibility studies — parsed, chunked, indexed, and cited. When you ask a question, the system returns the answer and the page it came from. No guessing.

RAG done right: frugal, precise, contextual. We don’t dump 100 pages of noise into the model. We retrieve only what’s relevant. Lower latency. Higher accuracy. Zero hallucinated fluff.

Spatial intelligence is built into the foundation. Real estate is geographic, and so is analysis. Competition clusters, trade areas, and demand patterns render directly on the map — the way we actually reason about location.

Insight flows through the actual workflow. Slack integration means risks, findings, and comments follow the deal automatically. Context lives with the team, not in someone’s inbox.

The result: analyst leverage grounded in verifiable data, not AI-generated prose. When AI is grounded in data, its insight becomes indisputable.