AI Startup Hebbia Raises $130 Million for AI-driven Complex Query Tool

TMTPOST--Hebbia, a startup leveraging generative AI to search large documents and answer complex questions, has secured $130 million in a Series B funding round.

The round of financing, led by Andreessen Horowitz and followed by Index Ventures, Google Ventures, and Peter Thiel, values Hebbia at approximately $700 million.

The company plans to use the funding to refine its tool, which uses artificial intelligence (AI) to search through documents and answer complex queries.

“Designed for the knowledge worker, Hebbia lets you instruct AI agents to complete tasks exactly the way you do them — no task too complex, no dataset too large, and with full flexibility and transparency of a spreadsheet,” the company said.

The announcement provides examples of how Hebbia’s product has helped its clients. For example, during the Silicon Valley Bank crisis, asset managers were able to sift through millions of documents to determine exposure to regional banks.

“It can execute complex workflows, not just chat back and forth,” said George Sivulka, the founder of the company.

Founded in 2020, Hebbia has developed software capable of analyzing a diverse array of digitized documents and data sources, such as regulatory filings, PDFs, and even audio and video clips.

The company's technology allows users to handle more complex queries than those typically managed by consumer-facing chatbots.

Hebbia's granular approach to displaying results helps users verify the accuracy of responses, addressing corporate concerns about the potential for AI to generate incorrect information.

This AI-driven complex query tool is already being used by a wide range of clients, including asset managers, law firms, banks, Fortune 100 companies, and the U.S. Air Force.

During the recent Silicon Valley Bank (SVB) crisis, Hebbia's technology helped asset managers map their exposure to regional banks by analyzing millions of documents, according to the startup.

Activist investors also used the tool to identify inconsistencies across more than 8,000 regulatory filings, a task that would be challenging for humans alone.

Corporate lawyers use Hebbia's software to gain a negotiation advantage by summarizing "market" terms in real-time.

Additionally, thousands of ad hoc AI agents have been employed to price private assets more effectively, conduct new types of due diligence, and screen a greater number of opportunities, Hebbia reported.

"Hebbia raises $130m for AI-driven complex query software" was originally created and published by Verdict, a GlobalData-owned brand.

Hebbia initially found success in the financial sector, where firms spend $100 billion annually, with teams dedicating over 60 hours per week to researching high-risk transaction data. Hebbia quickly secured paying clients among the world's largest private equity firms, hedge funds, and consulting companies.

Hebbia's capabilities extend to any knowledge-intensive industry. Management consulting firms can better understand client businesses through reports, and lawyers can review thousands of contracts for due diligence. By increasing the efficiency of these professionals by 20%, Hebbia can save millions and significantly reduce financial risks.

"Hebbia is tackling a multi-billion-dollar problem that has plagued industries for years: how to efficiently leverage vast amounts of sensitive data. The answer lies in Hebbia's technology. George and the Hebbia team are changing the way we find answers," said Mike Volpi, partner at Index Ventures.

Hebbia's flagship product Matrix is an AI agent capable of completing end-to-end tasks across multiple file types and formats, delivering structured answers with references. The system is transparent, showing the sources and steps taken to reach each conclusion.

By enabling models to handle any quantity and type of data while providing relevant references, Hebbia helps users track every action and understand how conclusions are reached. Its multi-modal and multi-model capabilities allow it to work with various models and data types, making it versatile for any information type and format.