The “marriage” of knowledge and enterprise AI

3 innovation and digital news in 1 minute. Every Monday. Episode 361

Pryon gathers 100m for knowledge x AI

“For those who can’t afford to wait for the right answer.” — this is the promise of newcomer Pryon, accessing knowledge database in enterprises. They gathered $160 million for that, very recently $100 million. A software that searches unstructured data in an enterprise with a large gain in time wasting. Opinion: Pryon develops a vocabulary of its own. They “weave” between enterprise data, fight against “knowledge friction” and build a “knowledge fabric” for that. Good vocabulary and yes, AI x knowledge is under-developed in any enterprise.

The next big thing?

Forgotten among the LLM noise? Cohere

Cohere, from Toronto, focuses on enterprise, mainly for a knowledge retrieval use case. They are one of the few with an own LLM, called Command. Today with a chat demo, called Coral. Cohere is behind Oracle, Jasper and Notion solutions. Opinion: Cohere was strong in the news in spring with a large funding of $435m and a founder who was part of the scientific team that wrote the “birth certificate” of transformer at Google Brain in 2017. Today the noise is so loud around Large Language Models that Cohere seems very quiet.

Quiet, but very well positioned, enterprise AI.

The name says it all — Weight and Biases

Weight and Biases is a software tool to tell and visualise where the AI has weights and which biases might occur. The condition for developing enterprise AI solutions do that in knowing what the AI does. Opinion: A great name that says it all, and courageous to name it that way. Let’s be precise, this is how they describe themselves: “developer-first MLOps platform that builds machine learning performance visualization tools”. Gathered a serious $250m for that.

Biases in AI — a challenge for dev teams.