Deasie Gathers $2.9M Funding to Promote Data Governance in AI

Tech startup Deasie has recently confirmed a fruitful $2.9 million seed funding round. This capital injection will help promote the company's expansion and development of their innovative toolkit aimed at affording tech companies better command over text-generating artificial intelligence (AI) models. Investment was raised from a pool of leading investors, including Y Combinator, General Catalyst, RTP Global, Rebel Fund, and J12 Ventures.

Focused on revolutionizing data governance and stewardship, Deasie brings together a trio of experienced founders-Reece Griffiths, Mikko Peiponen, and Leo Platzer. This startup group has previously worked on data administration tools at McKinsey. Noticing considerable issues around enterprise data governance, they comprehended the influencing factors and solutions that could effectuate the adoption of generative AI in companies.

In terms of large language models (LLM), like OpenAI's GPT-4, Deasie is striving to augment their dependability by creating a system that links unstructured company data such as files, reports, and messages. The aim is to automate their classification based on what they contain and their degree of sensitivity. The system uses pre-set tags and labels to organize and specify data, an approach that informs Deasie's algorithms on how to categorize future data.

Deasie has the functionality to automatically tag a report as containing 'personally identifiable' or 'proprietary' information, highlighting third-party versions of reports, restricting access rights and more. Subsequent to the auto-tagging process, Deasie scans through each tag to examine the corresponding data for its relevance and significance. This analysis guides the decision-making of which data to ingest into a text-generating model.

"Enormous amounts of unstructured data exist in enterprises, often overlooked from a governance perspective," explains Griffiths. Through their novel approach to data management, Deasie serves to elevate the security and quality of data for generative AI applications. The platform sifts through thousands of documents per company, ensuring the information selected is high-quality, pertinent and safe for use.

Existing products in the market take a 'data safety' or 'data governance for structured data' approach to LLM governance. Deasie fills the gap for a method of assessing data quality and relevance for unstructured data. By honing in on this unmet need, the company is stepping up to resolve the challenge of aligning every generative AI use case with the most compatible set of data.

This unique approach places Deasie in competition with established names like Unstructured.io, Scale AI, Collibra, and Alation.

Looking ahead, Deasie has plans to boost its engineering team and their recruiting efforts to carve their niche in the world of technologically advanced data governance, making the most of their recent funding success.