How marketers are actually building the data infrastructure to scale AI

AI dominated corporate earnings calls this year, with 65% of S&P 500 companies mentioning the technology during their Q1 earnings reports, a record according to FactSet. But while executives are eager to talk about AI, far fewer have built the data infrastructure needed to put it to work. Increasingly, executives say, that is the real challenge.

It all starts with data, Snowflake CMO Denise Persson told The Current. “You cannot have an AI strategy if you don’t have a data strategy first.”

That’s becoming a common refrain across the industry. With AI advancing at breakneck speed, marketers are confronting the less glamorous work of organizing, connecting and governing the data those systems depend on. In that sense, infrastructure itself is becoming a competitive advantage.

The IAB recently projected that ad spend would grow 9.5% year over year in 2026, largely due to the increasing use of agentic AI to speed up campaign planning, activation and optimization. Still, only half of marketers surveyed by the IAB say they’re currently scaling AI within their companies today, with most of those teams using it for analytics rather than planning.

Snowflake is seeing an explosion in customers looking to strengthen their data infrastructure as agentic AI envelops business plans, Persson said. Its latest agentic product aims to address two of the biggest barriers to enterprise AI adoption — enabling agents to work across different platforms and models while giving companies control and visibility into what those agents are doing.

Companies across the industry told The Current they’re betting that while AI products will evolve rapidly, investments in high-quality data will outlast any particular technology or model.

HyphaMetrics is one of those companies. For its first six years, the company focused on dashboards. Now, co-founder and CEO Joanna Drews said its product road map looks vastly different than it did just six to eight months ago.

While brands need a platform like Snowflake to organize first-party data, they also need reliable measurement data coming back. Without that, companies could be working off incomplete or misleading signals.

“How can you tell an AI agent what to do without the data of where people are spending their time?” Drews told The Current. “None of it is going to work without somebody fixing the measurement problem. That’s where it’s going to break down. That’s what’s going to stunt its growth.”

PadSquad underwent a similar shift in November 2024. But rather than creating a chatbot, the company focused on building an AI-powered feedback loop that helps its brands and agencies continuously optimize creative performance. That required a hard look at its own data.

“What we learned early, early on, to be honest, was our data was not organized in a way where it was something that AI could use in a meaningful direction,” Lance Wolder, the senior vice president of commercial strategy and marketing, told The Current.

Nearly two years later, PadSquad created the foundation needed to scale its offering. That included bigger things like connecting third-party signals and integrating its partners in a privacy-conscious way to something as basic, but critical, as standardizing naming conventions.

A piece of advice has stayed with Wolder for years: “Find the simplest, repeatable thing that has to be done day in and day out and use AI to solve that first. And then move to the next one.” That approach, he said, minimizes the desire to be pulled toward every new trend.

LiveRamp has also rolled out a series of agentic AI products in recent months, including a partnership with Nvidia focused on AI infrastructure. Matthew Hogg, senior vice president of business development, at LiveRamp stressed the importance of setting the right foundation for a company.

“We’re developing the right standards, the right governance, the right type of observability and support so that you can audit what’s being done and how it was done and why it was done in order to unlock long-term investment and growth,” Hogg told The Current in July. “[Then] it becomes this genuinely transformative thing, where in a few years, a lot of these teams and businesses are operating in a different manner than they are today.”