Eugene Mymrin/Moment via Getty Images Follow ZDNET: Add us as a preferred source on Google. ZDNET's key takeaways Snowflake now offers MCP support. The company also debuted an AI suite for financial institutions. Both aim to make agentic AI more secure for sensitive datasets. On Thursday, data platform Snowflake announcde its own Model Context Protocol (MCP) Server, in keeping with the quick trend Anthropic set with the protocol last spring. The company also announced a new package of AI tools specifically geared toward financial services, allowing them to develop AI agents safely and securely. Also: How Snowflake's new tools turn business analysts into AI developers The new tools are an extension of Snowflake's Cortex AI, the company's existing AI services bundle that includes agents, data querying features, analysis, and partnerships with frontier models from Anthropic and OpenAI, among other perks. Cortex AI for Financial Services is tailored for that industry. Also: ChatGPT can now connect to MCP servers - here's how, and what to watch for AI systems, especially agents, are only as effective as the data to which they have access. Financial institutions have sensitive proprietary data about clients, trades, and transactions -- keeping that secure while also leveraging it for automating tasks, fraud detection, analysis, and more can be a delicate process. "Embedding AI in financial services requires a dual approach: making the technology deeply understand the business and making it easily accessible to everyone while maintaining trust and security," Snowflake said in a blog post. According to a recent MIT study, 95% of organizations struggle to see ROI from AI adoption, running into issues like mistrust, poor application, insufficient training, and lack of security infrastructure as roadblocks. MCP Server In short, MCP allows LLMs and agents to securely connect with all kinds of previously siloed, proprietary, or otherwise inaccessible data sources. In public preview, the server lets businesses connect Snowflake AI with external agents. For example, Snowflake has integrated Anthropic's Claude for Financial Services with Cortex AI. Also: Anthropic now lets developers use Claude Code with any remote MCP server "With Snowflake MCP Server, remote agents can now connect with Snowflake data -- as well as third-party data shares from Snowflake Marketplace through Cortex Knowledge Extensions -- enabling interoperability with the broader AI ecosystem," Snowflake said. For finance, the integration could help a portfolio manager connect the dots between news, market data, and earnings calls to find investment opportunities. It could also help a banker finalize lending decisions using summarized risk intelligence from property data, customer conversations, and loan documents, according to Snowflake. Also: Snowflake launches Openflow to help businesses manage data in the age of AI Financial institutions can also use Snowflake MCP Server to connect to platforms and products from companies including Anthropic, MistralAI, Cursor, Salesforce's Agentforce, Workday, and more. Cortex AI for Financial Services Snowflake announced a set of tools for integrating AI agents with finance. According to Snowflake, the new suite makes tasks like claims management and quantitative research easier (and scalable) with AI while keeping data pathways secure. Also: Snowflake's new AI agents make it easier for businesses to make sense of their data The tools bring together insights from an organization's proprietary data, external structured data, and information from unstructured data like journalistic articles for "richer, more context-driven insights," Snowflake said in a blog. The company's goal is to provide an end-to-end pipeline for businesses that preps their data for AI, applies agentic solutions, and creates observability, all while maintaining governance and security standards. "Financial services firms rely on data scientists for risk modeling, forecasting, trading analytics, and compliance, but much of their time is spent on data preparation and repetitive coding," Snowflake noted in the blog. The new tools build on the company's existing AI offerings, including Snowflake Intelligence and Data Science Agent, as well as Cortex AISQL, which lets analysts use SQL query language commands to contain what AI models generate, making the information easier to contextualize. Also: Too many AI tools? This platform manages them all in one place For financial services, Snowflake Openflow helps businesses port data like call center transcripts and claims reports across silos, while Cortex AISQL lets analysts transcribe earnings calls, structuring previously hard-to-use data into a workable tool. Cortex Agents, in public preview, help glean insights from structured and unstructured data as well. Data Science Agent uses natural language to help institutions build credit risk models or fraud detection systems "in minutes," Snowflake said. "Automating data cleaning, feature engineering, model prototyping, and validation so teams can move from raw data to production-ready models faster." With features like Snowflake Marketplace, users can also access third-party data securely to help automate workflows. Cortex AI for Financial Services then allows clients to connect directly and securely with multiple frontier models. How to try it Cortex AI for Financial Services is available today. Snowflake MCP Server is in public preview starting today. Enterprises can request a demo or contact Snowflake sales directly. Also: 50 AI agents get their first annual performance review - 6 lessons learned Get the morning's top stories in your inbox each day with our Tech Today newsletter.