TL;DR:
- Deasie secures $2.9 million in seed funding with investments from Y Combinator, General Catalyst, RTP Global, Rebel Fund, and J12 Ventures.
- The startup, founded by McKinsey alumni, addresses the need for improved data governance in generative AI.
- A recent IDC survey underscores the demand for enhanced governance to ensure the quality of generative AI insights.
- Deasie’s solution automatically categorizes unstructured company data, making it reliable for generative AI models.
- The platform allows users to define data tags and labels, shaping Deasie’s algorithms for future data classification.
- Deasie’s technology evaluates data relevance and importance, ensuring high-quality input for text-generating models.
- Despite having just three employees, Deasie secures a pilot agreement with a “multi-billion-dollar” U.S. enterprise and has a robust pipeline of enterprise customers.
- Deasie distinguishes itself by measuring data quality and relevance for unstructured data, setting it apart from competitors.
- The company plans to expand its engineering team and introduce new features to stand out in the market.
Main AI News:
In the realm of cutting-edge AI innovation, Deasie emerges as a formidable player, poised to revolutionize the landscape of generative AI. With a recent infusion of $2.9 million in seed funding from key investors such as Y Combinator, General Catalyst, RTP Global, Rebel Fund, and J12 Ventures, Deasie is on a mission to empower businesses with enhanced control over text-generating AI models.
Deasie’s visionary founders, Reece Griffiths, Mikko Peiponen, and Leo Platzer, bring a wealth of experience from their tenure at McKinsey, where they identified critical challenges and opportunities in enterprise data governance. These insights led them to recognize the profound impact data governance could have on the successful integration of generative AI within organizations.
Their observations resonate with the broader industry sentiment. A recent IDC survey of over 900 executives at major enterprises revealed that a staggering 86% acknowledge the pressing need for more robust governance to ensure the “quality and integrity” of generative AI insights. However, only 30% of respondents expressed confidence in their readiness to harness generative AI’s potential today.
In response to these challenges, Deasie has developed an ingenious solution that targets the reliability of generative AI models, particularly large language models like OpenAI’s GPT-4. The heart of their product lies in its ability to seamlessly connect with unstructured company data, including documents, reports, and emails. Deasie’s innovative technology automatically categorizes this data based on its content and sensitivity.
For instance, it can swiftly tag a report as “personally identifiable information” or “proprietary information” while indicating its version number. Alternatively, it may label a spec sheet as “proprietary information” and emphasize its restricted access rights. Crucially, Deasie allows its customers to define these tags and labels according to their specific data classification needs. This collaborative approach essentially “teaches” Deasie’s algorithms to accurately classify future data.
Once the documents are tagged, Deasie’s platform delves into its library of tags to assess the overall relevance and importance of the corresponding data. Based on this evaluation, it makes informed decisions about which data to channel into a text-generating model.
Reece Griffiths, one of Deasie’s co-founders, emphasizes the significance of their endeavor: “Enterprises have enormous volumes of unstructured data that have rarely received any attention from a governance perspective. The probability that language models retrieve answers that don’t make sense or are exposed to sensitive information, scales with the volume of data. Deasie is an intelligent platform that filters through thousands of documents across an enterprise and ensures that data being fed into generative AI applications is relevant, high-quality, and safe to use.”
The concept behind Deasie is undeniably intriguing. The idea of curating an LLM’s data sources to vetted, trusted information is a commendable approach, particularly considering the risks associated with unleashing language models on outdated or contradictory data. However, questions linger about the consistency of Deasie’s algorithms in classifying data and the potential for occasional misjudgments regarding a document’s importance.
Nonetheless, Deasie appears to be gaining traction in the market. Despite having only three employees, the company has secured an agreement for its first pilot with a “multi-billion-dollar” U.S.-based enterprise. Additionally, it boasts a substantial pipeline of over 30 enterprise customers, including five Fortune 500 companies.
Deasie takes a unique stance in the realm of generative AI governance. While other products focus either on “data safety” or “data governance for structured data” in LLM governance, Deasie stands out by addressing the critical issue of measuring data quality and relevance for unstructured data. They have pioneered innovative approaches in this domain.
In the coming months, Deasie has ambitious plans to expand its engineering team and make multiple strategic hires. Their primary focus will be on developing features that set them apart from competitors such as Unstructured.io, Scale AI, Collibra, and Alation. As Deasie continues to evolve, it promises to be a driving force behind the responsible and effective adoption of generative AI in the corporate world.
Conclusion:
Deasie’s successful funding round and innovative data governance solution address a critical need in the market. The demand for reliable generative AI is evident from the high percentage of executives recognizing the necessity for enhanced governance. Deasie’s platform has the potential to reshape the landscape by providing a structured approach to data quality and relevance, enhancing the adoption of generative AI in enterprises.
