Google has invested heavily in generative AI, the most recent AI surge, as well as other areas outside of its own exclusive technologies. Alphabet, the parent firm, is also investing in the startup field — AlphaSense.
AlphaSense, a supplier of business data, received a $100 million investment from CapitalG, the late-stage venture capital branch of Alphabet, according to CNBC. This investment valued AlphaSense at $1.8 billion.
When it comes to offering statistics on businesses that can aid shape investment and corporate strategy, AlphaSense competes alongside firms like FactSet and Bloomberg.
The fundraising phase, which has been revealed on Tuesday, comes after months of growing excitement around generative AI, especially OpenAI’s ChatGPT and other text-generating programs which use large language models, or LLMs, to offer unique and complicated replies to user queries. Google announced Bard, a conversational tool in February that will operate with both its main search engine as well as other applications.
According to AlphaSense CEO Jack Kokko, the startup is creating a function that would intelligently simplify financial information for users so they can more conveniently retrieve the most vital information. Summarization has always been difficult work for AI technologies, but with LLMs’ assistance, it has improved vastly.
AlphaSense was developed in 2011 by Kokko. After a $225 million investment in June that’s been managed by Goldman Sachs and Viking Global Investments, the most recent fundraising is a flat round. The value from a 2021 financing led by similar investors doubled during this round.
Because the phrase hadn’t yet reached the common language, generative AI wasn’t a subject for discussion in the previous two sessions. In a news release from last year, AlphaSense claimed that their platform “extract relevant insights from an extensive universe of public and private content” using “proprietary search technology” backed by AI and natural language processing.
Since the end of 2021, publicly-traded company values have deteriorated sharply, as well as the Initial Public Offering sector has been frozen. The only field of hope for the year has been generative AI, which has grown somewhat frothy in some quarters.
A two-year-old pre-revenue firm named Character launched in March. In a funding round led by Andreessen Horowitz, AI—founded by two former Google employees—raised $150 million at a $1 billion capitalization.
With $100 million in yearly recurring profit already reached by AlphaSense in 2022, the firm is considerably further ahead. According to Kokko, the new funding is going to be utilized to hire more salespeople as the firm gears up to go public after the market gets stabilised. He stated that by using advancements in generative AI, the funding will also support AlphaSense in improving its technologies.
The use of relatively modern LLMs by AlphaSense, according to CapitalG partner James Luo, can enhance the primary product’s attractiveness to clients who don’t typically use financial services. If the user interface was simplified, salespeople might be more inclined to use a tool like AlphaSense.
“These are the people who are using Google Search to try to find every piece of information but they don’t have access to a lot of proprietary content,” Luo said of potential new users. “If you don’t work in that world, you need something that makes it a lot easier for you to understand that information.”
However, current LLMs encounter an issue known as “hallucination,” which refers to the inclination of the software to produce misleading data.
Kokko stated that AlphaSense is developing ways of making sure that its technology generates appropriate summaries with document references so that users are aware of the information’s origins.
He did not disclose which LLMs in specific AlphaSense intended to use, but he did say that the firm is analyzing nearly all of the important ones that are on the market at the moment. LLMs have been developed by firms like OpenAI and Cohere in addition to tech giants like Alphabet and Meta.