Closewell

Inspiration

We were inspired by the rise of AI copilots in professional services—tools like Harvey for lawyers and Copilot for developers. But in real estate, agents still face a frustrating, manual grind of managing documents, catching inconsistencies, and answering repetitive client questions. We thought: what if realtors had an AI copilot too?

That’s where Closewell began.

We wanted to build something that doesn’t just read documents—it understands them. Something that empowers agents to close deals smarter and faster, while also elevating the experience for clients.

What it does

Closewell is an AI copilot for real estate agents that helps them manage and understand the documents involved in closing a deal. It has three core functions:

Smart Document Completion & Consistency Checks It ensures that all required legal documents—like lease agreements, disclosures, and purchase contracts—are filled out completely and consistently, reducing human error and saving time.

Insight Extraction from Supplemental Documents Realtors can upload documents such as HOA meeting minutes or inspection reports, and Closewell will automatically summarize key insights—like pest issues, upcoming zoning changes, or other red flags that could affect property value.

Interactive Q&A for Agents and Clients Both realtors and their clients can ask questions directly to the documents—such as “Can this property be sublet?” or “What’s the HOA fee?”—and receive clear, accurate answers in natural language.

How we built it

We powered Closewell using Gemini, leveraging its strengths in document understanding, summarization, and natural language interaction. Our backend uses Gemini for Q&A and insight extraction, while the frontend delivers a clean, unified interface for document uploads, auto-completion, and client interaction. This allowed us to build a real estate-specific AI copilot quickly while maintaining high accuracy and usability.

Challenges we ran into

One of our biggest challenges was designing an AI experience that fits naturally into a realtor’s workflow. Real estate agents aren’t legal experts or technologists—they need fast, actionable support without extra complexity.

We also struggled with how best to surface insights from unstructured documents, like HOA meeting minutes, and present them in a way that’s actually useful during a sale. Fine-tuning Gemini’s responses to stay relevant and concise took careful prompt engineering.

Finally, UI design was a challenge—we needed to balance simplicity for everyday use with enough depth to support complex document review, while also making it intuitive for clients to interact with the assistant.