Milena Traikovich is a powerhouse in the demand generation space, specializing in how brands can turn complex data into high-quality lead-nurturing machines. In this discussion, we explore the tectonic shifts happening at the intersection of data warehousing and artificial intelligence, specifically focusing on the recent developments from the Snowflake Summit ’26. We delve into how the elimination of data silos, the integration of foundational models like Claude, and the rise of agentic AI are reshaping the marketing landscape, allowing teams to move faster while maintaining ironclad governance and data privacy.
How does the strategic shift of bringing AI models directly to the data warehouse change the way marketing teams manage sensitive customer information and security risks?
The traditional workflow of exporting massive datasets to external AI platforms has always been a point of friction and a major security vulnerability for any performance-driven team. By integrating Anthropic’s Claude models directly into the Snowflake environment, we are seeing a massive reduction in the need to move sensitive customer information across different systems where it might be exposed. This “System of Intelligence” approach allows us to perform complex sentiment analysis and trend exploration while keeping the data governed under one roof. Marketers can now focus on generating high-quality content and deep audience insights without the constant fear of data leaks or compliance violations that usually occur during the export process. It truly changes the game by making privacy a built-in feature of the creative process rather than a secondary hurdle to clear at the end of a project.
With the launch of Cortex Sense as a context layer, how do you see AI agents evolving to handle the specific nuances of a company’s internal language and performance metrics?
The biggest frustration with general AI models has always been their tendency to hallucinate or lack the specific context of a unique brand’s campaign structure or product catalog. Cortex Sense acts as a crucial context layer that translates a company’s specific business rules and internal terminology so the AI can actually understand the environment it is working in. This means that instead of a generic answer, a marketing team gets insights that respect their specific audience definitions and performance KPIs without needing to provide constant manual guidance or “hand-holding.” We are moving toward a reality where agents built with tools like CoWork and CoCo can autonomously navigate complex workflows because they finally have the “tribal knowledge” that was previously locked in the heads of senior strategists. It allows for a level of reliability in AI-powered workflows that we simply haven’t seen in the enterprise space until this push into agentic AI.
In what ways does the ability to share AI agents across different accounts rethink the traditional, often siloed, relationship between brands and their marketing agencies?
The introduction of Cortex Agent Sharing is a significant move toward transparency and efficiency because it allows brands to share the “intelligence” of an agent without ever exposing the underlying raw data. For instance, a brand could grant an agency access to a specialized audience analysis agent that has been trained on their specific customer behavior, yet the agency never actually sees the sensitive database records or PII. This setup, combined with expanded support for Apache Iceberg and open data architectures, eliminates the need for endless copying of datasets and the creation of redundant data silos that typically plague large-scale campaigns. By working from a single, governed source of truth, both the brand and the agency can stay perfectly aligned on execution while maintaining the highest standards of data architecture. It fosters a much more collaborative and secure ecosystem where the focus remains on campaign optimization rather than the plumbing of data transfer.
How does translating complex governance and privacy rules into plain English through tools like the Horizon Catalog change the dynamic between marketing and technical data teams?
For too long, data governance was viewed as a back-office concern that primarily functioned as a bottleneck for agile marketing teams trying to launch new initiatives. By moving toward a conversational treatment of governance where access rules are defined in plain English, we are empowering marketers to take an active role in privacy and compliance. These updates allow the Horizon Catalog to automatically turn simple instructions into enforceable policies across all data, AI tools, and agents, ensuring that every customer-facing workflow stays within approved rules. This speed is essential because as AI moves directly into the front-end of the business, we cannot afford to wait weeks for a technical team to manually set permissions for every new experiment. It essentially bridges the gap between the people who understand the customer journey and the people who protect the data, creating a more cohesive and responsive organization that doesn’t have to choose between security and speed.
What is your forecast for the evolution of marketing workflows as these integrated systems of intelligence become the industry standard?
I anticipate a future where the “stitching together” of dozens of disparate marketing tools becomes a relic of the past as we embrace a more unified, data-centric architecture. We will likely see a massive shift where marketing professionals spend far less time on data preparation and manual movement, allowing them to focus almost exclusively on high-level strategy and creative activation across the entire customer journey. The emergence of agentic AI that lives exactly where the data resides will mean that real-time personalization and complex trend analysis will finally be achievable at scale without the latency or security risks we face today. Ultimately, the successful marketers of the next decade will be those who can orchestrate these internal AI agents to manage the customer experience with built-in governance and unmatched data quality from the very start.
