Pryon Inc. today announced the company raised $100 million in a Series B round led by billionaire Thomas Tull’s U.S. Innovative Technology Fund to accelerate the growth of its platform, which helps enterprises build artificial intelligence assistants that can ingest and understand their vast stores of unstructured data and other knowledge.

The company was founded in 2017 by Pryon Chief Executive Igor Jablokov who comes from an interesting AI pedigree. He originally created the company Yap, a Siri-like speech recognition startup that Amazon.com Inc. acquired in 2011 to begin developing its talking assistant, Alexa. Before launching Yap, Jablokov also worked as the lead for the multimodal AI research team at IBM Corp.

Pryon produces an AI-driven platform that allows companies to quickly get their “knowledge” into a format that can be easily understood by a chatbot or an AI-powered search interface so employees and customers can use it to surface insights. To do this, Pryon trains and deploys its own proprietary models that are selected for the content needed that are grounded to avoid hallucinations – the tendency for AI models to make errors — and are grounded in the specific knowledge.

“My lifelong goal, and Pryon’s mission, is to reduce the distance between knowledge and people, especially for those who make up the backbone of critical infrastructure,” said Jablokov. “Our platform brings organizations to a state of knowledge flow, empowering a decision advantage grounded in confidential, trusted and verifiable information.”

The company’s platform is capable of ingesting from a large number of sources including audio, images, text and video files. Enterprise data can come from a variety of sources, including web pages, annual reports, marketing presentation, and the platform supports numerous file types including Microsoft PowerPoint, Google Docs, Slides, text files, PDFs, secure connections and more.

Pryon also uses optical character recognition and computer vision to allow companies to ingest large amounts of text and graphical data from paper manuals and books, which means that even if the data isn’t electronically stored, it can still be pulled into a format from which the AI can understand and pull insights. This can include handwritten documents and forms produced in the field. For video and audio, the platform can use speech recognition to transform spoken words into knowledge graphs for the AI.

The company said that it has specifically trained a number of large language model AIs for business formats that can quickly ingest for any industry content such as energy, financial, healthcare, industrial, technology or utility. This allows companies across every market and sector to quickly bring their knowledge into a format that an AI can understand and have it at their fingertips.

Once a model is trained by ingesting knowledge from all of these sources using enterprise data, users can use it using natural language conversational requests in text or voice. It can be connected to enterprise apps through a large number of interfaces including a chatbot style or via search.

The funding round included existing investors and new participation from Aperture Venture Capital, BootstrapLabs, Breyer Capital, Duke Capital Partners, Good Growth Capital, Omnimed Capital, Revolution’s Rise of the Rest Seed Fund.

Since its launch, Pryon has gained a number of big-name clients across large enterprise markets, including Dell Technologies Inc., Nvidia Corp. and Westinghouse Electric Corp.

Jablokov said that the new funding will go toward hiring new employees to expand the sales team and build out the current platform and product. The company also plans to increase outreach into international markets as well as grow strategic partnerships.

Image: Pixabay