Pryon, a startup focused on developing an AI-powered platform for extracting insights and surfacing answers from enterprise knowledge databases, has announced the successful completion of a $100 million funding round, with Thomas Tull’s U.S. Innovative Technology Fund leading the investment.
Founder Igor Jablokov stated that the newly acquired capital will be directed towards supporting Pryon’s overall expansion, growing its team of 100 employees, increasing its global presence, and expanding its strategic partnerships. A source with knowledge of the matter informed TechCrunch that this funding round, which brings Pryon’s total funding to $137 million, values the company at an estimated post-money valuation ranging between $500 million and $750 million.
Before founding Pryon, Igor Jablokov led IBM’s multimodal AI research team. He later left to establish Yap, a speech recognition startup similar to Siri, which was acquired by Amazon in 2011 to accelerate the development of Alexa. (Interesting tidbit: Pryon’s name comes from Amazon’s code name for the speech engine powering Alexa.)
While Pryon is not a voice assistant, it serves as an assistant in its own right. Jablokov characterizes it as a “knowledge fabric” capable of interfacing with third-party chatbots or channels. It ingests various data types, including audio, images, text, and video, and transforms them into a searchable and usable format compatible with the connected frontend.
An analogous service to Pryon is Amazon’s Kendra, which relies on AI and machine learning to facilitate enterprise search. Like Kendra, Pryon employs connectors to unify and index previously disparate data sources from databases. However, Jablokov asserts that Pryon surpasses Kendra in accuracy by up to 2x, ingests data up to 10x faster, and can index billions of documents, in contrast to Kendra’s limit of 100,000 documents.
Jablokov emphasizes that organizations do not need to migrate their content into the Pryon platform. Instead, it overlays existing systems of record and does not necessitate retraining end-users to author content differently. It simply points to a repository and generates an AI model from the underlying content. Pryon uses computer vision, optical character recognition, and handwriting recognition to understand the content, even if it includes legacy materials.
Pryon can create, update, or delete content on the platform in less than a second while maintaining privacy. Importantly, it leaves no trace of its indexing work. Jablokov highlights that customers have control over what goes into Pryon, whether it’s public, published, proprietary, or personal data. This ensures attribution to authorship and ownership, preventing unauthorized content from being included.
Pryon faces competition from Kendra and Microsoft SharePoint Syntex, both of which leverage knowledge bases to provide answers to company-specific questions. Additionally, startups like Hebbia, Kagi, Andi, and Glean utilize machine learning models to deliver specific content in response to queries rather than simple lists of results.
Despite the competition, Pryon has achieved significant success, with annual recurring revenue reaching the “seven figures” mark and securing “a dozen” large enterprise and public sector clients, including Dell, Nvidia, and Westinghouse.
Jablokov underscores Pryon’s unique suitability for enterprise use since it was designed with AI capabilities from its inception. The platform can meet the stringent requirements of highly regulated environments, from the energy sector to government, thanks to its innovative content safeguards.
