Data governance platform, Deasie has raised $2.9 million in a funding round closed just five months post-launch. European-founded by Reece Griffiths, Mikko Peiponen and Leo Platzer, and US-headquartered, the startup supports the adoption of language model applications.
The Seed funding was backed by Y Combinator, General Catalyst, RTP Global, Rebel Fund and J12 Ventures. The fresh funding will see Deasie launch a talent search for senior engineers to expand its product team.
Deasie’s platform connects to a company’s data sources compiling all data into smaller sections based on semantic meaning, tagging it for contents and sensitivity. The library of data can be used in a GenAI application, like a chatbot - the language model is first filtered for relevance, quality and safety.
“Nearly 80% of today’s data is unstructured - think documents, reports, text, images. With the rise of GenAI, companies are now attempting to leverage these to feed into chatbots, knowledge assistants, content-generation tools etc,” says Griffiths.
“However, the historic lack of quality controls and governance around this type of data means that AI applications are being fed with low-quality and sensitive data, leading to the deployment of unreliable and unsafe models. We want to change that by allowing companies to intelligently vet the data they are feeding into these AI models," he says.
“The founding team behind Deasie have the needed deep domain expertise in this area to help enterprises take full advantage of LLMs. I look forward to becoming more involved with them as they scale,” says Nicolas Dessaigne, Partner at Y Combinator.
Lead image: Deasie founders. Image: Uncredited.