Real estate transactions are notoriously paperwork-heavy. Brokers and their teams spend hours pulling property records, tax documents, and deeds, then manually filling contracts and managing the operational steps of every deal. This administrative burden slows down closings and pulls brokers away from what matters most — working with clients.
RealPact is building voice-powered AI agents that handle this operational work automatically.
As a Y Combinator Summer 2026 company founded by Dartmouth students, RealPact is an early but focused vertical AI agent play targeting one of the most document-intensive workflows in real estate.
Data
Funding Stage: YC S26-backed with additional support from LeapYear Fund, Dartmouth Venture Studio, and Real Estate Angels. The company has raised approximately $500K+ and is in the process of raising a larger round.
Launch / Founding Date: Founded 2025. YC Summer 2026 batch with early traction in New Hampshire and plans for rapid expansion.
Key Leadership:
Ranvir Deshmukh, Co-Founder — Dartmouth student and key driver behind the product vision.** Erik Peterson**, Co-Founder — Dartmouth student who, together with Ranvir, won the 2026 Magnuson Startup Competition.
Team size is currently 3, based in San Francisco (with strong ties to Dartmouth’s entrepreneurial ecosystem).
Core Tech Stack / Approach: Voice-powered AI agents designed specifically for real estate brokerage operations. The system automatically retrieves deed, property, and tax records, auto-fills contracts and brokerage forms, and manages transaction workflows. It emphasizes a voice-native interface for ease of use and focuses on deep integration with real estate data sources (MLS, county records, etc.). The architecture is built as an “AI-native operating system” for brokerages rather than a general-purpose tool.
Editorial
Plain English Pitch (2 sentences): RealPact gives real estate brokers voice-powered AI agents that automatically pull property records, tax documents, and deeds, then fill out contracts and manage the steps of a transaction. Instead of spending hours on paperwork, brokers can speak naturally and let the agents handle the operational work so they can focus on clients and closing deals faster.
ICP & Primary Use Cases:
Primary buyers are real estate brokerages and individual agents (starting in New Hampshire and expanding nationally) who are burdened by the administrative side of transactions. These teams deal with repetitive, time-consuming tasks like gathering records and filling forms that slow down deals and create operational drag.
The core problem solved is the heavy paperwork and workflow management burden in real estate transactions, which reduces broker productivity and extends deal timelines.
Key use cases include automatic retrieval of property/deed/tax records, auto-filling of contracts and brokerage forms, voice-driven transaction management, and moving deals from initiation to close with significantly less manual effort.
Hiring Patterns:
As a small team of three in the YC S26 batch, RealPact is in active early growth mode. Expect focused hiring in engineering (especially AI agents and real estate data integrations), sales, customer success, and operations as they expand beyond initial New Hampshire customers and build out the platform.
Buying Signals:
- YC Summer 2026 acceptance.
- Early revenue from brokerages in New Hampshire with plans for national expansion.
- Strong investor interest from real estate-specific backers (LeapYear, Real Estate Angels) alongside YC.
- Winning the Magnuson Startup Competition at Dartmouth.
- Clear product traction with agents already handling live transaction workflows.
These signals indicate both product-market fit in a specific vertical and readiness to scale.
Proprietary Insights
Proprietary Score — Real Estate Transaction AI Agent Index:
RealPact scores very highly on this custom early-stage metric. Key strengths include the founders’ deep focus on real estate transaction workflows, the practical voice-powered interface tailored for busy brokers, early customer adoption in New Hampshire, YC validation, and backing from real estate-specific investors. The combination of vertical domain expertise and agentic automation for a notoriously manual process gives it strong positioning in proptech.
Competitor Matrix (Editorial Comparison):
| Dimension | RealPact (Voice-Powered AI Agents for Transactions) | Traditional Transaction Management Software | Manual Broker Workflows | General AI Tools / Assistants | Other Proptech Automation |
|---|---|---|---|---|---|
| Core Strength | End-to-end transaction automation with voice interface | Record-keeping and basic workflow | Full human control | Generic assistance | Niche tools |
| Automation Depth | High (records + contract filling + workflow) | Medium | Low | Low to Medium | Variable |
| Ease of Use for Brokers | Very High (voice-native) | Medium | High (but slow) | Medium | Variable |
| Vertical Specificity | Very High (real estate transactions) | High | N/A | Low | Medium |
| Current Stage | YC S26, early customers in NH | Mature | Ubiquitous | Broad | Growing |
| Best For | Brokerages wanting to reduce admin time dramatically | Compliance and record management | Small volume or complex deals | General productivity | Specific pain points |
Founder & Company Vision Highlights:
Ranvir Deshmukh and Erik Peterson are building RealPact with the belief that real estate software should handle the operational work behind every deal. Their vision is to create an AI-native operating system for brokerages where voice-powered agents retrieve records, draft contracts, move transactions forward, and manage workflows — freeing brokers to spend more time with clients instead of buried in paperwork.
Deeper proprietary perspectives on the product roadmap, specific data source integrations, expansion strategy beyond New Hampshire, and long-term vision for AI in real estate operations are best gathered through direct conversations with the founding team.
Why This Matters in 2026
Real estate remains one of the most document- and process-intensive industries, yet much of the operational work is still handled manually. Vertical AI agents that deeply understand real estate transactions and can execute them with minimal human intervention have the potential to meaningfully improve speed, reduce errors, and free up broker time. RealPact is one of the earliest dedicated attempts to build this layer specifically for brokerages.