Skip to content

Data Room Auto-Builder

Data Room Auto-Builder is a Gemini 3.5 Flash multi-agent system that reconstructs a startup founder's investor data room from scattered sources in under a minute.

Code Video Cerebral Valley

Core trick

Five sub-agents pull documents from connected accounts, classify them, and diff the result against a diligence checklist to list gaps.

Tags

financelegal

Description

Data Room Auto-Builder is a Gemini 3.5 Flash multi-agent system that reconstructs a startup founder's investor data room from scattered sources in under a minute. Five sub-agents (Scouts, Classifier, Extractor, Gap Analyzer, Synthesizer) pull documents from Gmail, Drive, Carta, Stripe, Ramp, and Notion, tag and extract metadata, compare findings against a Series A diligence checklist, and produce an organized data room plus a red flag report of what's missing. It solves a real founder pain: when a term sheet lands, docs are scattered everywhere and founders either delay closing or pay a law firm $15k to $30k to assemble the data room under pressure. This tool compresses six hours of doc archaeology into thirty seconds
AINative-Studio/google-io-hackathon-data-room★ 00 forksPythonCreated May 23, 2026 · last push Jun 1, 2026

README excerpt

# Data Room Reconstructor

**Google I/O Hackathon 2025 — Multi-Agent Founder Data Room Reconstruction**

> Turn 30+ hours of scattered founder docs into a 100% investor-ready data room in seconds — powered by Gemini agents running in parallel, with automatic gap detection, AI-generated fixes, and live export to OpenCap Stack, Carta & Pulley.

---

## Demo Video

https://github.com/AINative-Studio/google-io-hackathon-data-room/blob/main/demo.mov

> Watch the full agent swarm run: 4 parallel scouts → gap analysis → auto-fix → live push to OpenCap Stack

---

## The Problem

Founders raising capital scatter critical documents across 4–6 different systems:

| Source | What Lives There |
|--------|-----------------|
| Gmail | Invoices, tax docs, term sheets, investor emails |
| Google Drive | Financial models, board minutes, cap tables |
| Carta | Equity grants, option pool, shareholder records |
| Stripe / Ramp | MRR, ACV, burn rate, expense breakdowns |

VCs expect a **complete, organized data room** before any serious diligence conversation. Founders spend **30–40 hours** manually hunting, organizing, and formatting this data — and still miss critical documents that kill deals.

---

## The Solution

A **4-phase multi-agent system** built on Gemini (via AINative) that:

1. **Scouts** all data sources in parallel
2. **Identifies** every gap against the full investor due diligence checklist (63 documents across 8 categories)
3. **Auto-fixes** gaps by generating investor-grade doc

Team