Lovelace AI Inc. is launching a new approach to enterprise artificial intelligence, focusing on high-stakes decision-making environments where errors have significant consequences.
The core of their offering is Elemental, a "context engine" builder situated between AI agents and data systems. Elemental generates secure, enterprise-specific context engines that transform fragmented data into structured knowledge graphs. These graphs enable AI agents to navigate and query information, delivering research-quality analysis with citations.
Elemental's backend, YottaGraph, can manage trillions of interconnected facts, augmenting internal data with external intelligence. The system uses a fraction of the token use compared to other methods, allowing for more extensive querying.
Founded by Andrew Moore, former head of AI at Google Cloud and dean at Carnegie Mellon University, Lovelace AI targets industries like public sector, national security, disaster response, and healthcare, as well as financial services.
Lovelace AI is ingesting billions of facts weekly from diverse public sources. A key feature is its handling of data relationships, processing millions of messages between nodes in the knowledge graph for efficient computation. The system ensures accuracy through entity resolution and maintains trust by tracking provenance for every inference, providing citations for the entire reasoning process.
The platform is deployed within customer environments, offering crucial control for sensitive applications. Lovelace AI will focus on large enterprises seeking tangible AI-driven productivity gains, especially in areas where AI implementations have historically faced challenges.