Snowflake launches Cortex Agents to enhance AI data retrieval and decision-making for enterprises
In addition to Cortex Agents, the firm has also introduced Cortex Analyst, which uses Anthropic Claude for text-to-SQL queries to facilitate high-quality structured data retrieval.
American AI data cloud company has expanded its AI capabilities with the launch of Cortex Agents.
The new offering extends Snowflake’s existing Cortex AI retrieval services, where the agent streamlines agentic app data access and orchestration for enterprises to make AI-driven decisions.
According to Snowflake, Cortex Agents are designed to orchestrate data retrieval across both structured sources (such as Snowflake tables) and unstructured sources (like PDF files in object storage), enabling enterprises to quickly derive actionable insights.
Data retrieval refers to the process of obtaining relevant information from various storage systems and formats, ranging from database tables to PDFs or text documents.
“At Snowflake, our strategy is to provide AI for our customers in an easy, efficient, and trusted manner. All of our customers start with data at the core and then build on top of it,” says Baris Gultekin, Head of AI at Snowflake, in an interaction with YourStory.
“Snowflake has built a premier data platform to be able to analyse large amounts of data and governance. With AI, our customers can do a lot more,” he adds.
In addition to Cortex Agents, the firm has also introduced Cortex Analyst, which uses Anthropic's Claude for text-to-SQL queries to facilitate high-quality structured data retrieval. Cortex Search, another key offering, focuses on unstructured data, claiming accuracy that outperforms 's embedding models by at least 11% on several benchmarks.
AI agents are autonomous systems that perform complex and multi-step tasks in minutes to boost enterprise productivity. However, for agents to be effective, they must access both structured and unstructured data. As data connections multiply, managing access controls and retrieving accurate information while maintaining privacy becomes challenging.
The quality of agent outputs relies on the data and the accuracy of retrieval systems. Many organisations struggle with an AI-data mismatch. While large language models excel at handling unstructured data, many lack mature data preparation practices. Cortex Agents address this need by streamlining the integration, retrieval, and processing of diverse data types, enabling customers to build high-quality, scalable data agents.
For example, a financial analyst might need to combine structured revenue data with unstructured financial reports and market research. Such enterprise scenarios demand secure access and governance to ensure the right information is surfaced to AI systems.
“We are seeing a lot of adoption across the board. Everyone’s building super exciting chatbots, where documentation that used to be hard to analyse is now suddenly available,” says Gultekin.
“We are also seeing strong interest in using AI for business intelligence. Today, companies have to go through a long waiting time to unlock data that's already sitting in business accounts. For instance, a lot of our customers have dashboards that they look at to understand product trends, revenue. However, these dashboards tend to be relatively rigid. Once they want to dig into the next level of details, they need to go to an analyst and the analyst needs to write reports — that takes time and effort. Cortex Analyst makes it very easy for businesses to just use natural language to query such business intelligence data,” he adds.
Several Snowflake customers, including Siemens Energy, Bayer, and Nissan, are already using Cortex Agents.
The agent deconstructs complex queries, retrieves relevant data, and generates precise answers using Cortex Search, Cortex Analyst, and LLMs, ensuring accuracy and governance throughout the process.
“At Luminate, we're revolutionising how we deliver precise, data-driven insights to our clients through generative AI applications. Snowflake's unified data and AI platform provides our developer team with scalable processing and retrieval for both structured and unstructured data — the critical building blocks for developing, deploying and orchestrating data agents powering our application,” said Glenn Walker, Chief Data Officer, Luminate Data.
Snowflake has also integrated Anthropic’s latest Claude 3.5 Sonnet model, delivering advanced reasoning, coding, and workflow execution under the same governance standards that apply to data. This model powers Cortex Analyst and is available for use within Cortex Agents for improved structured data analysis, and enhanced agentic workflows.
Gultekin adds that India will play a key role for Snowflake going forward.
"India has long been at the forefront of technology adoption. We see India as a strategic market for Snowflake and already have many valued customers there. With its young and dynamic market, I believe India will adopt AI faster than average," says Baris.
Edited by Jyoti Narayan