Highlights:

  • The company claims that its software is used by investment firms managing USD 14 trillion in assets to help them make informed financial decisions.
  • According to Hebbia, it generates analyses by combining large language models with computer vision models.

Hebbia Inc. raised USD 130 million in the investment spearheaded by Andreessen Horowitz. The company develops data search platforms for businesses.

Peter Thiel, Alphabet Inc.’s GV venture fund, and Index Ventures were also involved. According to a source, Hebbia’s current net worth is estimated to be around USD 700 million. The investment was somewhat expected since late June when reports indicated the company was raising capital.

Using natural language prompts, employees can sort through massive volumes of data with the help of the platform that bears their name. The company claims that its software is used by investment firms managing USD 14 trillion in assets to help them make informed financial decisions. Lawyers use the software to summarize extensive legal papers.

The software arranges search results in a manner akin to a spreadsheet. In response to a user’s inquiry regarding the financial performance of the top five competitors in a particular market, the platform may generate a distinct spreadsheet column for each company. The columns for each corporation can then be divided into fields that show information about revenue, net profit, and earnings per share.

Hebbia can retrieve and analyze corporate data. For instance, a financial analyst could ask the platform to rate the top ten investment possibilities according to feasible returns. Real-estate professionals can utilize Hebbia to produce property assessments.

Under the hood, the platform uses internal company records, regulatory filings, charts, and other files to respond to employee requests. According to Hebbia, it generates analyses by combining large language models with computer vision models. The software has a feature called a routing mechanism that searches for the best neural network to handle a user’s query automatically.

Some LLMs find it challenging to perform tasks that require several phases. Hebbia claims that its platform tackles this problem by dissecting intricate user requests into discrete steps and matching each step to a distinct neural network. The software uses LLMs from OpenAI to perform some processing functions.

In conjunction with a recent investment announcement, Hebbia said that, from an undisclosed base, its revenue had climbed by a factor of 15 over the previous 18 months. During the same period, the employment of the company had increased fivefold. With the additional cash, Hebbia will be better equipped to sustain its expansion in the face of growing competition from other AI firms.

In February, Glean Technologies Inc. raised USD 200 million for its competitor business search platform. The program can establish a connection with an organization’s internal applications and enable natural language prompts for searching the documents they contain. Additionally, Glean has a programming assistant that can provide developers with code guidance.