Here are the most influential voices across industries that IT professionals should know.
Matt is an associate editorial director and award-winning content creation leader. He is a regular contributor to the CDW Tech Magazines and frequently writes about data analytics, software, storage and cloud topics.
Matt is an associate editorial director and award-winning content creation leader. He is a regular contributor to the CDW Tech Magazines and frequently writes about data analytics, software, storage and cloud topics.
Snowflake is a powerful tool to help organizations take advantage of artificial intelligence by bringing structured and unstructured data together in a single environment and making it easier for organizations to prepare data for AI models.
But before they can achieve the AI success that Snowflake offers, organizations need to use it effectively, said Paul Zajdel and Christopher Marcolis of CDW during a presentation this week at Snowflake Summit 26 in San Francisco.
“A lot of companies say they haven’t been able to take it to the next level,” Zajdel, CDW’s vice president of data and analytics, said. “You have to have certain fundamental elements in place to take advantage of all of Snowflake’s features.”
The results can be spectacular for organizations that use AI effectively. “We’re seeing tremendous change in the world when it comes to Snowflake and what it can do,” said Marcolis, head of presales for data with CDW.
To get there, organizations must pay attention to several key details such as infrastructure, data governance and semantics.
Click the banner below to get the latest BizTech insights in you inbox.
The first step is to build a foundation that includes infrastructure elements such as storage and compute, enabling an organization to move and use data effectively. Governance is critical at this level, Zajdel said, but is often overlooked due to the pressure many business leaders feel to show business value quickly in their AI initiatives. Not only should organizations establish strong governance at the outset, but they should maintain that effort, he added.
“Data governance is not a project. It doesn’t have a beginning or an end,” Zajdel said. “It’s a way of life. You have to commit to it.”
Strong governance enables organizations to put guardrails on AI so that users can trust the outcomes it produces, Marcolis said. This is especially important when AI agents are used to create additional agents.
“You have to have this foundation in place, or you’re not going to realize the full potential of what AI can be,” Zajdel added.
Next, the transformation layer enables organizations to govern the outcomes of its AI projects and tokenize the data it ingests. At this layer, some organizations have implemented a “medallion architecture” that organizes data into progressive layers of quality and refinement.
The third layer, the semantic layer, makes analytical AI more explainable, consistent and trustworthy at scale by bringing semantics closer to the data and defining common business concepts so outputs such as dashboards, reports and AI agents produce consistent answers.
DISCOVER: Turn data into insights and accelerate artificial intelligence initiatives.
At the intelligence layer, AI becomes operational. “The robots are taking over, and you have to be ready for that,” Marcolis said, noting that AI is useful when the lower layers are disciplined and governed.
Many organizations begin their AI projects at this step, which leads to problems. By building their intelligence on a governed platform, organizations can achieve targeted outcomes, Marcolis said.
The next step is the agency layer. At this point, humans may begin to go beyond asking agents to achieve outcomes and instead have agents create new agents to achieve these outcomes.
Paul Zajdel Vice President and GM, Data & Analytics, CDW
“We have to think about giving autonomy to our data,” Marcolis said, but he noted that many organizations are reluctant to take this step.
Finally, the commerce layer is where organizations can use the data they have to bring in new revenue streams. Marcolis cited the example of a water filtration company that gleans valuable health insights from the particles it filters out of a water supply and then sells that data to organizations that were never part of its original customer set.
Achieving this level brings clear value, but organizations have to set up every other layer effectively to get there. “The companies that have done the work are going to have a significant advantage when they start using data,” he said.
READ MORE: How can AI agents support lean security teams?
Many executives are reluctant to begin using AI because they don’t know where to start, Zajdel said, but this is a mistake that demonstrates “paralysis by analysis.”
Organizations that follow the correct steps will be able to turn their most creative and innovative ideas into reality. “Those that wait are going to get left behind,” Marcolis said.
Zajdel’s guidance was simple and clear: Organizations that haven’t already put these steps in place should get started now.
“You’ve got to start somewhere,” he said. “Just start small and build from there.”
Keep this page bookmarked for our complete coverage of Snowflake Summit 26.
Learn how IT leaders are working to build a frictionless enterprise.
Security
5 Questions About Security Debt for Small Businesses
Artificial Intelligence
Cisco Live 2026: AI Could Help Security Teams Move Faster, Say Cisco and OpenAI Leaders
Visit Some Of Our Other Technology Websites:
Tap into practical IT advice from CDW experts
Copyright © 2026 CDW LLC 200 N. Milwaukee Avenue, Vernon Hills, IL 60061
Do Not Sell My Personal Information
