Accelerating enterprise AI use cases into production is now a board-level priority for most companies. However, one of the key challenges in AI today is ensuring that those use cases are ready for real-life use and continue to perform at a high level in production. Not only must enterprises ensure accurate, reliable, and valuable results they must also address and mitigate  critical issues like bias, hallucinations, and toxicity. The ability to demonstrate that AI is both trustworthy and high performing will be key to AI adoption and its ability to deliver ongoing business value.

At Snowflake, we are making massive investments in generative AI and end-to-end machine learning capabilities to help customers build and deploy high-impact AI use cases that maximize the value of their data. In particular, we have made several advancements to Snowflake Cortex AI, our fully-managed generative AI service, and Snowflake ML, our set of capabilities for training, deploying, and running predictive models. 

Today, we’re excited to announce a new, complementary investment that will allow us to provide even deeper functionality that will help organizations drive AI quality and trustworthiness by evaluating, monitoring, and debugging models and apps across the full lifecycle, in both development and production.

Snowflake has entered into a definitive agreement to acquire the TruEra AI observability platform, which provides leading capabilities to evaluate and monitor LLM apps and ML models in production. 

TruEra's technology helps evaluate the quality of inputs, outputs, and intermediate results of LLM apps. This expedites experiment evaluation for a wide variety of use cases, including question answering, summarization, retrieval-augmented generation-based applications (RAG apps), and agent-based applications. TruEra’s technology also provides detailed, actionable insights to improve ML model performance and accuracy by revealing anomalies in model metrics and providing a specific root cause analysis for rapid debugging. 

TruEra AI Observability also can help to identify LLM and AI risks such as hallucination, bias, or toxicity, so that issues can be addressed quickly, and so that organizations can demonstrate compliance with AI regulations.

TruEra’s capabilities complement the AI and ML data governance functionalities we already provide in the AI Data Cloud. Snowflake provides deeply integrated capabilities to ensure the accuracy and trustworthiness of the data used to supplement and train models, and the observability technologies developed by TruEra will complement and round out that story for AI.

We’re also excited to welcome many of TruEra’s talented engineers and executives who bring deep expertise in model observability and explainability. They include TruEra’s three co-founders: President and Chief Scientist Anupam Datta; Chief Technology Officer Shayak Sen, and CEO Will Uppington. 

We look forward to collaborating with the TruEra team to bring exciting new capabilities around AI and LLM observability to the AI Data Cloud in the future.