eWEEK content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
Snowflake has announced the integration of Anthropic’s Claude 3.5 generative AI into its Cortex Agents service, enhancing corporate data processing capabilities. This move aligns with the growing trend of AI-powered automation in enterprise data management.
The companies announced on Feb. 12 that Claude 3.5 will join the tools in the Cortex Agents service, which is now in public preview. This integration enables more advanced AI-powered data retrieval and analysis, allowing enterprises to work more effectively with structured and unstructured data.
Claude in Cortex Agents
Claude 3.5 powers two key tools: Cortex Analyst and Search. These tools enhanced structured and unstructured data retrieval capabilities, making AI-driven insights more accessible to businesses.
Cortex Agents (which includes Analyst and Search) is part of Cortex AI, Snowflake’s bid for a platform built on the lucrative AI market. Cortex AI lets organizations build AI tools (including LLMs, RAG and text-to-SQL services) based on the organizations’ own structured or unstructured data.
Cortex Agents adds AI functionality to application data access and orchestration, while Cortex Search performs unstructured data retrieval accuracy. Specifically, Claude in Cortex Agents can:
- Answer: Users’ specific and ambiguous questions about data.
- Split: Complex tasks into subtasks for better execution.
- Choose: The appropriate Snowflake product — Cortex Analyst, Cortex Search, or SQL generation — based on natural language questions.
Snowflake says that Claude-powered Cortex Search achieves 12% greater accuracy in unstructured data retrieval than OpenAI embedding models.
“LLMs excel at unstructured data, but many organizations lack mature preparation practices for this type of data; meanwhile, structured data is better managed, but challenges remain in enabling LLMs to understand rows and columns,” wrote Snowflake Senior Manager for Product in Generative AI and AWS Harshal Pimpalkhute and AI Product Marketer Arun Agarwal in a blog post.
Expanding the role of AI Agents
The addition of Claude 3.5 to Cortex Agents reflects the broader shift toward agentic AI — models capable of autonomously performing complex, multi-step tasks. Enterprise AI agents must balance automation, with compliance, ensuring privacy and security guidelines are met while retrieving and processing data accurately.
“Data agents, a specialized category of AI agents, combine data and tools to deliver more accurate, grounded insights by effectively selecting the right data sources and tools for retrieval,” Pimpalkute and Agarwal wrote.
Snowflake expects that customer support analysts, engineers, and other enterprise users will benefit from the enhanced productivity provided by Cortex Agents. Snowflake Head of AI Baris Gultekin noted in a media roundtable attended by VentureBeat that client application developers can also monitor what Agents are doing to refine answers and respond to governance or security needs.
Competitive landscape
Snowflake is in direct competition with other AI-powered enterprise data solutions:
Databricks: Offers a Data Intelligence Engine which uses AI to handle natural language queries.
Google Cloud BigQuery: Integrates with Google Gemini. Google’s AI model is for AI-powered data management.
Amazon Redshift: Employs AI for SQL recommendations, enabling users to generate queries from natural language prompts.
Criticisms of Agentic AI
Criticisms of generative AI add-ons or agentic AI include trustworthiness, lack of confidence due to unclear or missing regulations, and copyright concerns. Additionally, generative AI models require significantly more electricity and water than traditional computing, raising sustainability questions.