Snowflake’s Cortex AI features has been enriched to provide financial services companies with agentic artificial intelligence capabilities honed to their specific needs, the first of a planned suite of editions focused on individual industries.
Cortex AI for Financial Services will feature all the functionality of the platform’s Cortex features but will offer clients large language models that have been trained to respond to prompts suited to financial services companies’ workflows. In particular, the product seeks to address the critical bottleneck of extracting meaning and insights from data locked in a fragmented array of dashboards, spreadsheets and reports.“Think about all your data under one platform with governance, with controls and then effectively this really powerful enterprise intelligent agent that sits on top of that, allowing you to ask questions of your data across the enterprise – that, in a nutshell… is the value proposition here,” Rinesh Patel, Snowflake Global Head of Financial Services told Data Management Insight.
Data Nuances
The latest edition extends the Cortex AI framework for machine learning and natural language processing. General AI tools often struggle with the nuances of financial data such as the specific regulatory weight of Know Your Customer (KYC) data or the complexities of alternative investment research.Snowflake prioritised financial services in the development of Cortex because the largest share of its clients is from that industry, Patel said.
“It’s also an industry that is very, very data intensive,” he said. “Therefore, that was one of the reasons why it made a lot of sense to start there.”
The redesign focuses on three specific pillars:
Industry-Specific Semantic Layers: Transforming raw data into “AI-ready” data that speaks the language of financial markets.
The Managed MCP Server: ****A critical addition that enables secure interoperability with remote agents. This managed Model Context Protocol (MCP) server connects proprietary and third-party data from partners like FactSet, MSCI and Nasdaq eVestment directly to AI agent platforms.
Snowflake Intelligence: An enterprise intelligent agent specifically tuned to reason through front, middle, and back-office financial queries via natural language.
Fragmentation Crisis
Modern financial firms are drowning in data but starving for insights. As the “aperture of data broadens,” firms are managing a deluge of unstructured data covering market research, earnings call transcripts and transaction details that typically require manual review or complex ETL (Extract, Transform, Load) processes.
Snowflake has moved away from “model dominance” — the idea that one LLM rules all — and instead offers access to best-of-breed models. Within the Snowflake environment, users can use:
Diverse Model Selection: This includes proprietary models like Mistral, Anthropic’s Claude and OpenAI, alongside open-source models like Meta’s Llama series.
Snowflake Cortex AISQL: This tool adds functions for AI-powered extraction and transcription, allowing users to process documents, audio, and images at scale.
The “Agentic” Layer: Snowflake Intelligence selects the right model based on cost, speed and accuracy requirements.
“Depending on the cost, performance, the problem that you’re solving… the customer will choose the large language model of their choice and we handle the difficult challenges,” Patel said.
Cortex AI for Financial Services is designed to be deployed enterprise-wide, tackling specific challenges across different sub-sectors.
Asset managers can use the Snowflake Data Science Agent (currently in PrPr) to automate data cleaning and model prototyping, uncovering alpha in new alternative data sets such as those provided by The Associated Press or Investopedia.
Banks are deploying risk and compliance tools to optimise lending margins, reduce credit losses and automate KYC and AML workflows.
In the insurance sector, firms are using Cortex AISQL to analyse unstructured claims data and transaction details to speed up payouts and detect fraud.
Future Trends
The future of AI in finance is a move away from “the art of the possible” toward measurable commercial outcomes, said Patel. The industry is entering a phase where the success of an AI strategy is judged by its ROI. This requires a unified approach to data governance.
“The ROI is simply not going to be there if the data is not good, is not governed,” said Patel.
Looking ahead, the shift toward multi-agent architectures — where collaborative crews of specialised agents automate complex processes — will become commonplace.
Snowflake is betting that the future of finance lies in a conversational, governed and deeply integrated intelligence that turns data from a storage burden into a primary driver of growth.
“The biggest shift is really a pivot from a mindset change — an acceleration towards the commercial outcome rather than the technical win,” said Patel. “Customers are starting with the commercial outcome led by the business… and then lastly, they’re thinking about how the AI or the large language model can be the connective tissue between the data and the commercial outcome.”
Subscribe to our newsletter
