Prion, a startup developing AI-powered platforms to find insights and surface answers in enterprise knowledge bases, today announced it has raised $100 million in a funding round led by Thomas Tull’s US Innovative Technology Fund.
Prion’s founder, Igor Zablokov, said the new cash will be used to support Prion’s general growth, expand its 100-person team, expand its presence in international markets and enhance its strategic partnerships. A source familiar with the matter told TechCrunch that the funding, which brings Prion’s total raised to $137 million, values the company post-money at between $500 million and $750 million.
Before launching Prion, Jablokov led the multimodal AI research team at IBM. He left to create Siri-like speech recognition startup Yap, which Amazon acquired in 2011 to boost development of Alexa. (Fun fact: Prion was the code-name Amazon used for the speech engine that underpins Alexa.)
Priyan is not a voice assistant. but this Is An assistant – of sorts.
Jablokov described it as a “knowledge content” that can interface with a third-party chatbot or channel, ingesting data such as audio, images, text and video and converting it into a format that Which is discoverable and usable by any frontend connected to it.
An analog, Jablokov says, is Center, Amazon’s machine learning-powered service for AI and enterprise search. Similar to Centr, Prion leverages connectors to integrate and index previously disparate sources of information from databases. But Jablokov claims that Prion is 2 times more accurate than the center, ingests data 10 times faster and can index billions of documents compared to the center’s 100,000-document limit.
“Organizations do not need to migrate their content to the Prion platform because it layers onto existing systems of record and does not require retraining of the end-user to rewrite content,” Jablokov said. ” “You simply point to a repository and it generates an AI model from the underlying content. If you have old content in there, that’s OK, because Prion uses computer vision, optical character recognition, and handwriting recognition to understand what’s there.
Jablokov claims that Prion takes less than a second to create, update or delete content on the platform in a privacy-preserving manner – and the platform leaves no trace of its indexing work.
Jablokov claims, “Since the client defines what goes into Prion in terms of public, published, owned, and personal data, there is always attribution of authorship and ownership, so that only the content for which they are legally entitled is used.” Deserve,” claims Jablokov.
Prion has competition from the above-mentioned hubs as well as Microsoft SharePoint Syntax, which works on a knowledge base to link together answers to company-specific questions. Startups like Hebbia, Kagi, Andy, and Glenn also use machine learning models to return specific content in response to queries (as opposed to straight lists of results).
But Prion appears to be doing quite well for itself, recording annual recurring revenues in the “seven figures” and securing “a dozen” large enterprise and public sector customers, including Dell, Nvidia and Westinghouse. Is.
“Prion is one of the few AI-native companies that was designed for enterprise use from its founding days,” Jablokov said. “It can meet the needs of the most regulated environments, from energy to government, because of the platform’s unique way of securing content.”