SXSW: CMOs “all in” with AI; say PR and agencies matter more than ever

As artificial intelligence reshapes marketing, two top brand leaders made clear at South By Southwest that the future won’t be built by machines alone. Even as they invest heavily in AI, Lyft CMO Brian Irving and Samsung Electronics America’s Allison Stransky said that as large language models scan the internet for signals, the brands that win will be those backed by strong agency thinking, human creativity and high-quality earned media that teaches AI what to trust.

During the session “How CMOs are Rewriting the Rules of Connection,” both executives offered a detailed look at how AI is already embedded in their organizations, not as a side experiment, but as an operational layer touching everything from campaign development to customer experience.

Inside the AI Playbook: Agents, Experimentation and New Workflows
At Samsung, Stransky described a coordinated, enterprise-wide “all-in” approach to AI, beginning at the CEO level. The company has stood up a cross-functional AI task force — pulling in marketing, legal, HR, customer service and other teams — to evaluate AI use cases, vet tools and prioritize where AI can drive the most impact.

“We have an AI transformation team… and every function has what we call an AI crew,” she said, describing a structure designed to break down silos and accelerate adoption across the business.

That work is complemented by a test-and-learn mindset. Rather than committing to a single platform, Samsung is piloting tools, running controlled experiments and building internal capabilities while also partnering with external platforms.

At Lyft, Irving described a similarly hands-on approach, anchored in experimentation and internal tool development. One early step was creating a “safe space” for teams to explore AI without immediate performance pressure, including internal challenges that encouraged employees to test tools in real campaign scenarios.

For every dollar we spend on new AI tools, we need two to three dollars to invest in training, he said.

More recently, Lyft has begun embedding AI directly into its marketing workflows. Using tools like Google’s Gemini, Irving said teams are building agents that can synthesize inputs from meetings, documents and campaign materials to generate go-to-market strategies at speed.

“It can take all of those conversations… and develop go-to-market strategies,” he said, noting that human marketers still shape and refine the output. “There’s massive human intervention, but it’s not on the administrative aggregation.”

The result is a significant expansion of capacity. Tasks that once required manual coordination can now be completed in a fraction of the time, allowing teams to focus on higher-value decisions. “We’re all hoping that time saved in one place will be applied in another place,” Stransky said.

Other applications range from AI-assisted CRM optimization, including tools that generate and evaluate hundreds of email subject lines in minutes, to more advanced experimentation, such as AI-enabled brand asset creation and hyper-personalized content libraries.

But even as both CMOs described increasingly sophisticated use cases, neither suggested AI is replacing the core functions of marketing.

Agencies Aren’t Going Away — They’re Becoming More Critical
If anything, the complexity of the AI ecosystem is reinforcing the need for strong external partners.

“Our agency partners are still indispensable partners,” Stransky said, emphasizing that agencies are expected to evolve alongside brands by integrating AI into their own processes.

Irving pushed back more bluntly on the idea that AI could reduce reliance on agencies.

“I have peers who are like, ‘I’m getting rid of my agency because of AI.’ I think that that’s wrong,” he said. “You want external perspective on innovation and creativity.”

Creativity and Judgment Still Define the Work
“I don’t think AI is going be more creative than humans,” Stransky said late in the session, underscoring what both CMOs see as the enduring advantage of human-led marketing.

Irving added that beyond creativity, AI also lacks judgment — the ability to determine what is right for a brand, a moment or a specific audience.

“It’s not human versus machine,” he said. “It is human and machine.”

AI can generate options, accelerate production and expand reach. But it cannot decide what should ultimately represent a brand. That responsibility remains with the humans: marketers and, critically, with the agency partners helping shape the work.

PR and Earned Media “More Important Than Ever”
For PR professionals, the most important shift may be happening in how brands are discovered.

As search evolves and generative engines become a new layer of visibility, Stransky said the workload has effectively doubled.

“SEO is not over,” she said. “But here comes GEO.” And PR is a large part of how LLMs are trained through credible, high-quality content.

“PR is more important than ever,” Stransky said. “We need real, credible information all over the internet. High-quality journalism is needed more than ever.”

The reason is increasingly technical as news coverage, brand storytelling and third-party validation all contribute to how a company is represented in AI-generated answers.

“The ability to generate … content at scale is going to accelerate so fast,” Irving said, pointing to a coming wave of easy-to-manufacture low-quality AI material. To compete, brands must be “incredibly prolific with quality branded content” — PR content that not only reaches audiences, but helps define how AI systems understand a brand.