Snowflake World Tour Paris 2025 marked a strategic turning point for the cloud platform: shifting from a reactive data analysis solution to a truly unified, AI-native intelligence ecosystem.
At the heart of this transformation, two pillars emerge:
Snowflake Intelligence, a conversational natural language interface that allows business users to query their data without technical knowledge.
Cortex Agents, a framework opening the door to fully automated, AI-driven enterprise workflows.
Devoteam, named Partner of the Year 2025, presented demonstrations of Snowflake Intelligence and OpenFlow at the World Tour booth, along with a client roundtable dedicated to Data Products and the Modern Data Stack. This award recognises our involvement in various Snowflake projects—migrations, Data Platforms, AI & Data Products—conducted in close collaboration with Snowflake.
Snowflake World Tour 2025: Intelligence, Cortex Agents, and the Data-AI Revolution
Snowflake World Tour 2025 materialises the transition from a cloud data platform to a unified, AI-native intelligence ecosystem. The central theme is the shift from reactive data analysis to proactive AI-driven action. All of this is executed within a secure, governed enterprise framework that connects seamlessly. This event emphasised a unique mission: making the future of data, AI, and applications simple, efficient, and reliable.
A solid AI strategy is impossible without a coherent data strategy. In this context, Snowflake presented a series of innovations designed to manage the entire data lifecycle, from simplified ingestion and transformation to governed sharing and AI-driven activation. The Snowflake World Tour 2025 revealed a strategy aimed at democratizing AI for business users, providing developers with powerful tools to create agents, consolidating the modern data platform, and deepening strategic alliances, particularly with Microsoft.
Snowflake Intelligence and Cortex AI
At the forefront of these innovations are Snowflake Intelligence and Cortex AI. Snowflake Intelligence presents itself as the conversational “gateway” to the enterprise’s data assets. It allows non-technical users to interact with complex data in natural language. This relies on the underlying Cortex AI engine, a suite of managed services that integrates multimodal analysis, text-to-SQL generation, and semantic search directly into the platform’s core. For developers, Cortex Agents represent the building blocks for automating complex business processes, supported by a comprehensive lifecycle management and monitoring framework.
Integration with Microsoft Fabric
The strategic partnership with Microsoft has been significantly strengthened, with an emphasis on open standards and interoperability. The integration between Snowflake and Microsoft Fabric, built on Apache Iceberg, enables a “single copy of data” architecture. This eliminates silos and reduces data movement between the two platforms. This collaboration extends across the entire Azure ecosystem, including Azure OpenAI and Azure Machine Learning. Snowflake thus positions itself as an essential cross-cloud intelligence fabric.
dbt Projects
To support this AI-centric vision, Snowflake is modernising the foundational data platform. The introduction of dbt Projects on Snowflake natively integrates this industry-standard transformation tool into the platform, consolidating a key part of the data engineering workflow. This, combined with performance accelerators like Generation-2 Warehouses and migration tools like SnowConvert AI, aims to lower the barrier to entry and accelerate the transition to an AI-ready database.
Observability
Snowflake has made significant progress in observability and data quality. The expansion of Snowflake Trail and the launch of AI Observability in Cortex provide the necessary tools to monitor, diagnose, and govern AI applications in production, ensuring reliability and compliance.
Key Announcements and Feature Status from Snowflake World Tour 2025
| Feature | Status (October 2025) | Primary Impact |
| Snowflake Intelligence | Public Preview | Democratizes data access for business users, enabling self-service analysis via a natural language interface |
| Cortex Agents | General Availability (coming soon) | Enables developers to create and deploy AI agents that automate complex, multi-step business workflows |
| Cortex AISQL | Public Preview | Lowers barriers to AI development by allowing SQL-proficient teams to create multimodal analysis pipelines |
| dbt Projects on Snowflake | Public Preview | Centralises and simplifies data transformation pipelines, reducing TCO and architectural complexity |
| Microsoft Fabric Interoperability | General Availability / Preview | Eliminates data silos and duplication costs between major platforms, enabling “best-of-breed” engine choice |
| SnowConvert AI | General Availability (Free) | Reduces costs, time, and risks associated with migrating legacy data warehouses to Snowflake |
| AI Observability in Cortex | General Availability (coming soon) | Ensures trust, reliability, and governance of AI applications in production through no-code monitoring |
| Snowflake Trail Enhancements | General Availability | Provides unified, out-of-the-box observability across data pipelines, applications, and infrastructure |
| AI/ML in Data Clean Rooms | General Availability | Unlocks collaborative predictive analytics on sensitive data, creating a powerful data network effect |
| Snowflake OpenFlow | General Availability | Orchestrates and automates data ingestion from any source to Snowflake, combining connectors, transformations, and continuous loading into a unified, fully managed flow |
AI: Snowflake Intelligence and Cortex AI
Snowflake World Tour 2025 presented a unified AI foundation built on two pillars: Snowflake Intelligence (conversational interface) and Cortex AI (managed services), aimed at integrating AI into daily enterprise operations.
Querying Data in Natural Language with Snowflake Intelligence
This secure AI chatbot enables business users to query their data in natural language, eliminating the need for technical knowledge. It generates visualisations, traces data lineage, and suggests actions. Its strength lies in simultaneously querying structured and unstructured data for complex contextual analysis.
The major asset for the enterprise is the native integration of Snowflake security. Access controls, data masking, and governance are automatically applied, ensuring compliant deployment in regulated sectors.
Cortex AI
Cortex AI is the suite of managed AI/ML services that powers Snowflake Intelligence and enables developers to create custom AI applications. Its key components:
- Cortex AISQL: integrates AI into SQL to create multimodal pipelines (text, image, audio) with functions like AI_EXTRACT (table extraction in 29 languages)
- Cortex Analyst: converts natural language into precise SQL queries via semantic models that understand business context
- Cortex Search: hybrid search (semantic + keywords) on unstructured data, facilitating RAG applications
- Document AI: parses complex documents (PDFs) to extract structured information
- Secure LLM Access: gateway to leading LLMs (OpenAI, Anthropic, Meta, Mistral) with inference within Snowflake’s perimeter to protect sensitive data
Frédéric Adet’s Analysis
“Cortex AISQL unifies the analysis of structured and unstructured documents in the same pipeline with the same SQL commands, assisted by AI. Whether it’s tables, voicemail messages, or call center transcriptions, everything is processed uniformly. This drastically simplifies enterprise workflows. Where other platforms require a dozen different complex workflows, Cortex AISQL enables end-to-end business processes in a single pipeline. From document arrival to data extraction, to interpreting weak signals and executing the response.”
Accelerating Enterprise AI Adoption with Snowflake Intelligence
Snowflake’s two-tiered AI architecture aims to overcome obstacles to enterprise AI adoption.
Consumerization of Enterprise AI: Snowflake Intelligence replicates the success of consumer AI assistants by allowing users to “talk to their data” in natural language. This approach bypasses traditional bottlenecks (BI dashboards, analyst queues) and integrates analysis into the natural thought process of business users, expanding the market to make every employee a data consumer.
Unification of Structured/Unstructured Data: A Main Competitive Differentiator. Unlike siloed market solutions, Snowflake Intelligence combines Cortex Analyst (SQL generation) and Cortex Search (document search) to answer complex business questions that require simultaneous analysis of structured data (sales tables) and unstructured data (reports, emails, transcriptions) in a single, conversational query. This unified approach addresses a fundamental business challenge: obtaining comprehensive and contextually relevant answers.
Frédéric Adet’s Analysis
“Snowflake Intelligence really impresses me. It’s a way of encapsulating Snowflake’s various AI features that minimises the boundary between business and technical teams. This interface allows any company stakeholder to ask questions that go beyond classic BI: Snowflake Intelligence combines consumer LLMs and internal structured or unstructured content to respond, contextualise, and propose prescriptions.“
Cortex Agents
This developer framework enables the creation of automated, AI-driven enterprise applications. Cortex Agents aims to build an “agentic enterprise” with intelligent orchestration of complex workflows.
Cortex Agents offers a 4-step architecture mimicking the human analytical process:
- Automatic Orchestration: intent analysis, decomposition into subtasks, routing to appropriate tools
- Tool Usage: execution (Cortex Analyst for SQL, Cortex Search for documents, custom tools)
- Reflection: evaluation of results, iteration, or final synthesis
- Monitoring & Iteration: performance tracking, trace analysis, continuous improvement via user feedback
Agents are native Snowflake objects with conversational context (threads), customizable (instructions, tone, underlying LLM: Claude, GPT).
Developer Benefits
- Complete REST API: creation, management, programmatic execution; stateful (production) and stateless (prototyping) modes
- Custom Tools: extension via stored procedures/UDFs for external integrations (Salesforce, messaging services, proprietary models)
- Integrated Observability: detailed logs/traces of each step (LLM planning, tool calls, generated SQL) via Snowsight/API
- SDLC Support: environment management, change control, CI/CD integration (example: Simon Data deploys via Snowpark Container Services)
Integration with Microsoft Teams
This strategy allows Cortex Agents to be integrated where users already work—within Microsoft Teams and Copilot 365.
This integration (in preview since September 2025) enables users to engage in multi-turn conversational interactions with their Snowflake data directly within the Teams interface. It supports multi-agent scenarios and is designed to bring timely data insights directly into the flow of business collaboration and decision-making.
Frédéric Adet’s Analysis
“Cortex Agents offers companies a simple way to manage their business processes. In the coming months, we will see automated analysis of emails, commercial documents, procedures, and other enterprise content. Through workflows connecting drives (SharePoint, Google Drive) to all sources via OpenFlow, you can easily code processes and obtain automated business responses for large volumes. Where it took a week to analyse documents and produce a decision-support synthesis, Cortex Agents delivers relevant work in minutes, without human intervention.“
A Deepened Partnership with Microsoft
Snowflake World Tour 2025 significantly emphasised the partnership with Microsoft. This collaboration ensures seamless interoperability and long-term relevance within the enterprise cloud ecosystem. This announcement reveals a multifaceted integration covering data platforms, AI services, and business applications, all anchored in a shared commitment to open standards.
This partnership notably relies on interoperability between Snowflake and Microsoft Fabric via open standards, particularly Apache Iceberg. This “single copy of data” architecture allows organisations to store their data once in Microsoft OneLake and access it with both engines without duplication. OneLake automatically translates metadata between formats (Delta Lake/Iceberg), enabling teams to collaborate on a single source and choose the best engine according to their needs, thus eliminating redundant data movement.
Snowflake also offers broader integration within the Microsoft ecosystem:
- Azure OpenAI: secure access to models (GPT-4) from Snowflake, enabling RAG workflows without data leaving the secure perimeter
- Azure ML: unified MLOps cycle via Snowpark (data preparation, feature engineering) and AzureML (training, deployment, management)
- Azure Data Factory: mature orchestration of large-scale ingestion/transformation pipelines from Azure sources
- Power Platform & Power BI: bidirectional Snowflake-Dataverse connector (preview) for integration into Power Apps, Power Automate, and Dynamics 365; popular native Power BI connector offering optimal performance and security for self-service analysis
Frédéric Adet’s Analysis
“The partnership with Microsoft Fabric marks a strategic turning point. Where Snowflake and Microsoft were perceived as competitors a year ago, they are now associated with combining Microsoft’s enterprise daily environment (Active Directory, Teams, SharePoint) with Snowflake’s data power and data sharing capabilities. This interoperability, after AWS last year, confirms that Snowflake is now considered an essential partner in the data market.“
Modernising Foundations for an AI-Ready Data Stack
Snowflake World Tour 2025 clearly established that a successful AI strategy is inextricably linked to a modern, scalable, and well-governed data foundation. Snowflake’s announcements in this area focused on three key pillars:
- Accelerating migration from legacy systems
- Improving core platform performance to handle AI-scale workloads
- Fundamentally simplifying the data transformation process by natively integrating a key open-source standard into the platform
Migration and Performance
To address the problem of data trapped in legacy on-premise systems, Snowflake launches SnowConvert AI. This free solution automates the entire migration. The tool analyses existing code (Teradata, Oracle), converts it to Snowflake SQL, and generates validation tests, drastically reducing the effort, cost, and risk of large-scale migrations.
To handle modern AI and analytical workloads, Snowflake deploys two major innovations. Generation 2 Standard Warehouses offer 2.1x superior performance through hardware and software improvements. Adaptive Compute, an intelligent resource abstraction service, automatically routes queries to the optimal cluster, balancing performance and costs.
Finally, Snowflake OpenFlow streamlines ingestion by unifying the integration of diverse sources (batch, streaming, files) via a single interface with hundreds of pre-built connectors, thus consolidating the fragmented landscape of ingestion tools.
Data Clean Rooms
Snowflake Data Clean Rooms enable secure, privacy-respecting collaboration on data between multiple parties. The latest announcements focus on improving usability, expanding to new industries, and, most importantly, integrating AI and machine learning capabilities directly into the clean rooms environment.
2025 Enhancements
Simplified onboarding, improved interface, strengthened cross-cloud support. Major innovation (October 2025): inclusion of datasets with governance policies defined in other databases, enabling centralised governance and decentralised collaboration. New API developer role to facilitate integration into custom applications.
Sectoral Expansion
Beyond marketing/advertising (attribution, audience targeting), Data Clean Rooms prove promising in retail/CPG (joint planning, demand forecasting, and promotional impact measurement without customer data exposure) and healthcare (clinical research, HIPAA-compliant predictive analytics).
AI/ML Convergence
Direct integration of AI capabilities (Python templates, custom containers, chatbots) allows partners to jointly build and train sophisticated predictive models on combined data without mutual exposure.
Strategic Implications
Data Clean Rooms replace obsolete sharing methods (FTP, APIs) that have become risky under strict regulations. They position themselves as the “secure multiplayer mode” of the AI Data Cloud applicable to all industries. AI integration creates a defensible network effect where pooled, fragmented data generates superior models, attracting more partners in a virtuous circle where value grows exponentially.
Frédéric Adet’s Analysis
“Snowflake pioneered secure data sharing four years ago, a revolution enabling instant sharing between stakeholders without heavy projects. Today, Data Clean Rooms go further. Notably, They enable the exchange of large datasets from two companies for machine learning and data science, identifying reliable gain signals while fully preserving confidentiality and trade secrets. This is a major differentiator that only Snowflake architecture can offer today, applicable in all sectors—industry, aerospace, manufacturing—and particularly crucial for intercontinental collaborations subject to different regulations (GDPR, Patriot Act). Competitors will take years to catch up with this capability.“
Snowflake Intelligence and OpenFlow Demo
The demonstration at the Devoteam booth aimed to illustrate Snowflake’s business pragmatism through a complete real-time pipeline:
- Ingestion via OpenFlow (tool for ingesting structured and unstructured data, particularly in Change Data Capture)
- Data transformation with dbt Projects (managed version of DBT)
- Visualisation/monitoring with Streamlit
The emphasis was on Snowflake Intelligence, which deploys much faster than a traditional visualisation application while offering superior responses. However, unlike classic BI dashboards that simply display KPIs, Snowflake Intelligence provides an interpretation of next steps, contextual analysis, and pragmatic synthesis that explains observed trends, thus enabling the transformation of real-time data into concrete business actions.
Conclusion
The Snowflake World Tour Paris 2025 confirmed what many had sensed: the era of the simple data platform is over. Snowflake now establishes itself as the foundation of a unified enterprise intelligence ecosystem, where data, AI, and collaboration are one.
This vision is shared and realised daily by Devoteam with its clients: making artificial intelligence useful, governed, and actionable. Being named Partner of the Year 2025 is not an end in itself. It’s the starting point of a new phase: one where every company can leverage the power of Snowflake and AI to act faster, better, and together.
The adventure has only just begun.