As more enterprise teams move from experimenting with AI to putting it into practice, trust, context, and governance are becoming non-negotiable. Snowflake’s latest platform updates - paired with Secoda’s lineage, documentation, and observability tooling - offer a clear path to embedding AI into everyday workflows. Together, Snowflake and Secoda give organizations a faster path to operationalizing AI with the trust, context, and governance required from day one.

New AI features in Snowflake

At this year’s Snowflake Summit, the platform introduced one of its most expansive sets of AI features to date. These releases signal a focused direction: Snowflake is positioning itself as a starting place for implementing agentic AI, where intelligent agents work across structured and unstructured data, automate tasks, and support measurable outcomes, all within a governed environment.

The updates span multiple layers of the stack. New Data Agents let business users query data using natural language, with built-in lineage and explainability. Cortex AISQL adds multimodal capabilities such as document and image processing directly in SQL, removing the need for separate pipelines. Snowflake also introduced AI Observability tools to monitor accuracy and performance in production, and the AI Governance Gateway to manage permissions, usage, and budget controls across AI applications.

For developers and data scientists, Snowflake ML continues to evolve with tools that streamline model training and deployment. A standout is the new Data Science Agent, which uses prompts to generate end-to-end ML pipelines automatically — making advanced workflows more accessible without sacrificing governance.

These improvements reflect Snowflake’s serious move toward enterprise AI and agentic AI readiness. By building these capabilities directly into the platform, from natural language interfaces to governance controls, Snowflake is positioning itself as a core runtime for intelligent applications. They’re building the kind of AI that organizations can use in production and support enterprise-scale use cases.

Extending AI workflows with Model Context Protocol (MCP)

Snowflake’s introduction of the Model Context Protocol (MCP) expands the surface area for enterprise AI. MCP defines an open standard for how AI systems access structured metadata in a consistent, secure way. Secoda fully supports MCP and allows tools like Claude and Cursor to connect directly to the metadata catalog and access trusted metadata such as lineage, glossary terms, documentation, and SQL context, right where you work.

Because Secoda integrates with the full data stack, including Snowflake, BI platforms, orchestration tools, and more, teams can extend MCP-powered workflows beyond Snowflake alone. This makes it possible to apply governance, context, and quality checks across the entire ecosystem, not just within one tool.

By combining Snowflake’s AI infrastructure with Secoda’s metadata platform, teams can:

• Accelerate AI adoption by moving quickly from prototype to deployment, with governance and context already in place.
• Ensure responsible AI by adding lineage, documentation, and quality scoring that AI agents can reference during execution, essential for sensitive industries.
• Extend AI across the stack by making context portable — so agents and analytics tools can deliver consistent, explainable results across every layer of the data environment.

For organizations using Secoda, this enables faster and more confident AI adoption on top of Snowflake — with greater trust, richer context, and better alignment between tools.

Why Secoda + Snowflake work better together

For Secoda customers, this shift is an opportunity to adopt Snowflake’s AI features faster and with more confidence. As Snowflake continues to invest in the deployment of AI across structured and unstructured data, the role of metadata, lineage, quality, and data governance becomes more central. While Snowflake provides the infrastructure, Secoda provides the context and operational controls that help scale AI across both technical and business teams.

AI agents and workflows are only as reliable as the data and context behind them. To trust Snowflake’s AI outputs and use them in real business settings, teams need more than just compute and models.

They need to be able to:

• Trace exactly where the data came from
• Validate that inputs are high quality and production-ready
• Access documented context — including owners, definitions, and usage history
• Apply governance across the entire data environment, not just within Snowflake

Secoda supports this by providing full-stack lineage, data quality scoring, documentation, and access controls. Together, Secoda and Snowflake help organizations create AI that is not only powerful, but explainable, traceable, and ready for production use.

Secoda AI: Powering AI across the entire data ecosystem

As organizations move from experimenting with AI to deploying it across teams, metadata and governance become essential to ensure outputs are accurate, explainable, and trusted. Snowflake provides a strong foundation for AI workloads, with built-in governance for data and models inside the platform. Secoda AI builds on this by offering a metadata control plane that spans the full data environment — giving both technical and business users access to consistent context wherever AI is applied.

Snowflake’s AI tools are well-suited for reasoning over structured and unstructured data, generating SQL with Data Agents, and managing workflows using built-in governance features. But these capabilities are limited to data that lives inside Snowflake.

They don’t offer visibility into how upstream changes in dbt affect downstream dashboards, how definitions vary across teams, or how governance policies should apply across external tools.

Secoda fills this gap. It centralizes metadata and knowledge from Snowflake and the broader ecosystem — including transformation tools like dbt, BI platforms like Looker and Tableau, and orchestration systems like Airflow.

Engineering teams can also build on Secoda’s API-first lineage and automation framework, integrating it directly into their pipelines and workflows. This reduces the manual effort required to document datasets, models, and processes. Both business and technical teams benefit from real-time access to rich metadata, enabling AI agents, interfaces, and analytics to operate with the context they need.

At the core of this experience is Secoda AI, built to reason across this broader context and deliver complete, trustworthy answers:

• It uses a multi-agent architecture to parse user questions, retrieve metadata, and reason through workflows that span multiple tools.
• It references lineage paths, ownership, access policies, and historical documentation to ensure responses are grounded in accurate context, minimizing risks like hallucinated SQL.
• It adapts to schema changes and validates results against permission-aware access policies to ensure responses remain current and compliant.
• It supports multi-parameter filtering and complex search across the full metadata graph, handling questions such as “Which finance dashboards are using undocumented PII columns?”
• It is embedded across Secoda’s documentation, search, catalog browsing, and monitoring, enabling users to interact with AI in the flow of their existing work, including in tools like Slack and Chrome.
• It generates ad hoc visualizations for metrics and trends, helping users explore data quickly without needing to spin up new BI dashboards.

Snowflake provides a strong foundation for building and running AI workflows within its platform. Secoda extends that foundation by making AI more usable and transparent across the entire data environment. Together, they enable organizations to shift from tool-specific implementations to connected, governed AI that works across the business.

Want to see what Secoda AI looks like in action? Miinto, one of Europe’s fastest growing fashion marketplaces, uses Secoda AI to power automated documentation, simplify technical discovery, and support their customer success teams.

Final thoughts

As enterprise teams move from experimenting with AI to deploying it in production, trust, governance, and context become essential. Snowflake’s new AI capabilities give organizations a strong foundation for building intelligent applications within its platform. Features like Data Agents, Cortex AISQL, and the AI Governance Gateway make it easier to bring AI to data workflows inside Snowflake’s secure environment.

But production-ready AI requires more than platform-level capabilities. To answer questions across tools, enforce consistent policies, and understand how changes ripple through systems, teams need visibility that spans the full data stack.

Secoda complements Snowflake’s AI stack by providing a centralized metadata control plane and collaborative workspace that supports richer context, transparency, and collaboration. Paired with Snowflake, it gives teams a clearer path to scaling AI with the confidence that the right context and controls are in place. 

Get in touch to see how Secoda fits into your Snowflake-powered AI strategy.