data.world, purpose-built for the AI era since our inception, has officially rolled out our most important AI-related product to date: the AI Context Engine™. This product is redefining how businesses interact with and extract value from their structured data, offering unparalleled accuracy, explainability, and governance.
The challenge: LLMs and business context
Generative AI, especially large language models (LLMs) like OpenAI’s GPT-4 (and now the even more impressive GPT-4o), have made waves by allowing users to engage in natural language conversations with incredibly vast amounts of content online, including data and code. However, these models often falter when dealing with structured, internal business data, lacking the context to interpret complex business terms and relationships. This gap has hindered an organization’s ability to move from interesting experiments into fully-integrated, revenue-generating AI solutions.
The fix: Enter the AI Context Engine
Our AI Context Engine addresses these challenges head-on. data.world is built on top of a knowledge graph architecture (now with 76 patents), which integrates metadata, processes, policies, people, and systems into a language AI can understand. When organizations connect LLMs to their proprietary data and business context that’s stored in the knowledge graph, responses become more accurate and explainable. Research shows that LLMs are 4.2 times more accurate with a knowledge graph than those relying solely on SQL databases.
Real-world impact
Organizations are already reaping the benefits of our AI Context Engine. Team members can now ask complex questions like, “How does our organization define ‘current retention rate’, and how have these rates changed over the last five years?” In response, they’ll receive precise, contextually rich answers.
Vip Parmar, Global Head of Data at WPP, highlights the transformative power of our AI Context Engine: “Enabling all of our colleagues to interact with our data applications and services in a natural way unlocks our data for creative use, irrespective of skills and specialisms. The AI Context Engine bridges the gap from discovery to use, seamlessly and effectively.”
Cris Hadjez, Senior Director of Data Governance at Norwegian Cruise Lines, adds, “The AI Context Engine from data.world is powering the future of Norwegian Cruise Line by producing insights through GenAI that enable us to deliver a positive customer experience. We’re accelerating implementations and significantly improving trust and adoption of AI.”
Why the AI Context Engine is special
Our AI Context Engine isn’t just about chatting with internal data; it’s about providing explainability and higher accuracy in AI responses. Where other tools have struggled to leverage GenAI on top of unique business data, our AI Context Engine excels. With the API, starter kits, and plugins, organizations can easily integrate our AI Context Engine—and rich business context— into advanced AI applications and experiences. That means even people without technical know-how can have meaningful conversations with organizational data.
Moreover, our AI Context Engine’s robust governance capabilities mean that questions and answers are auditable, secure, and traceable. This streamlines the process of meeting business and regulatory requirements, showing how LLM responses are generated and what data sources were used.
Additionally, our AI Context Engine stands out with its industry-first optimization framework. This unique feature enhances LLM response accuracy through multivariate testing and continuous learning, eliminating the need for model training (or retraining). This out-of-the-box optimization not only saves organizations on computing costs but also ensures GenAI responses keep getting better and more accurate.
Final thoughts
The launch of our AI Context Engine marks a significant milestone in the journey towards smarter, more reliable AI-powered data interactions for all. As businesses increasingly adopt generative AI, we know the AI Context Engine will be essential in unlocking the full potential of data, worker productivity, and employee joy.
To experience the AI Context Engine firsthand, schedule a demo today.
Finally, I put a lot of love into this much longer article when we first announced our AI Content Engine on March 7th, where it places it in historical context (the most exciting time in the history of tech in my very long entrepreneurial career). And then, not surprisingly and shortly after, the main theme at TED2024 was AI, which I recapped in this article (Part Two of a two-part series on how AI is rapidly accelerating creativity and leading to a Renaissance 2.0) and also spoke about at our monthly customer townhall (all customers get access to that on data.world University).
