AI isn’t just shaping marketing, it’s reshaping decision-making itself. With the rise of intelligent agents making purchases on behalf of consumers, brands are being forced to rethink what relevance, trust, and value mean in the ecommerce ecosystem.
That was the focus of AI’s Next Act: Relevance, Trust, and the New Rules of Ecommerce, a fireside conversation featuring Claire Southey, Chief AI Officer at Rokt, and Libby Rodney, Chief Strategy Officer at The Harris Poll. Together, they explored the shifting dynamics of digital commerce, the emergence of agentic AI, and what it means to future-proof both brand and workforce in a real-time, AI-enabled world.
Agents are here, and they’re buying fast
What happens when a machine, not a person, becomes your customer?
According to Claire Southey, this shift is already underway. “Somewhere between 25 and 30% of ecommerce purchases by the year 2030 will be made by an AI agent,” she said, citing recent research from Gartner.
That acceleration compresses the window for brands to make an impression. “Your opportunity to engage with a purchasing decision, which might have been minutes long, hours long, days long—is now condensed to the hundred-odd milliseconds that it takes an agent to make a decision,” she said.
Libby Rodney backed it up with data: while only a few consumers today use agentic AI for purchases, nearly one in five Gen Z respondents say they’d trust AI agents to handle shopping for travel, beauty, and everyday household goods.
Low-consideration brands face high-speed challenges
As AI agents take over routine purchases, brands selling low-consideration products may be the first to feel the disruption.
“You’re almost abstracted away from the decision,” Southey explained. “You get a lot less time in front of the consumer, and that makes it really hard. Without that direct interaction, you're basically driving a complete data-driven process.”
And these agents don’t buy based on brand power or shelf placement. They evaluate products based on specs, values, and data. “They’re processing thousands of pages of information instantaneously,” she said. “If you’re a tissue brand, that means your product is being judged on things like sustainability, price, and return policies, not just a name on a box.”
Rodney posed the question: Will brands like these innovate or disappear? Southey’s answer was clear: Brands will have to become more consumer-centric than ever, even if they’re getting less feedback in return.
From planning to iterating: how to lead through change
As the speed of AI adoption increases, so does the pressure on teams to adapt.
“The era of episodic change is over,” said Southey. “We’re now in a state of continuous transformation.”
At Rokt, that means planning cycles have gotten shorter. “We’ve moved from annual planning to quarterly planning,” she said. “You can’t plan your way through this. You need to iterate through it.”
Hiring and upskilling are evolving, too. Southey explained that Rokt has eliminated degree requirements for many roles and focuses instead on adaptability and mindset. “The skills we value most today are soft skills. Adaptability. Comfort with ambiguity. Willingness to view failure as a growth opportunity.”
To assess that during hiring, her team uses experiential interviews and prioritizes traits like resilience, curiosity, and the ability to self-navigate. “These are more important than any technical credential,” she said.
The future is agentic, and data is the differentiator
In a world where AI agents make purchasing decisions, clean and actionable data becomes the competitive edge.
“There are only three things you really need to succeed in machine learning,” Southey said. “Electricity, infrastructure, and data.”
But there’s a catch. Most AI agents are managed by platforms that may cache product data or rely on outdated catalogs. That puts pressure on brands to ensure their data is not only high-quality, but surfaced in real time. “You can lose a lot of traffic very quickly if you’re not feeding accurate, accessible information into the system,” she said.
When asked whether this could unlock faster consumer feedback, Southey said yes, with caveats. “It depends on how much of that transaction is happening in real time versus cached data. But when it is real time, you can respond faster. Which means better business decisions, made sooner.”
Innovation and ethics can (and should) move together
A recurring theme throughout the week touched on a long-standing debate: does prioritizing AI ethics mean slowing innovation?
For Southey, it’s not an either-or situation. “It’s a false dichotomy,” she explained. “We can—and must—do both.”
She highlighted the risks of ignoring ethical considerations. “You might get to market faster, but without trust, growth stalls. At scale, integrity becomes non-negotiable.”
According to Southey, building that trust starts with three essentials: transparency, value exchange, and control. “People want to know how their data is used, what they get in return, and have the ability to remove it.”
At Rokt, those principles are built in. “We operate a closed ecosystem powered by first-party data. That means we deliver real-time relevance without compromising on trust.”
From shopping carts to self-driving cars
As the conversation wound down, Rodney asked a lighter question: if an AI agent could take one task off your plate, what would it be?
“I want my car to take itself to the dealership,” Southey said. “In the middle of the night, drive itself there, get serviced, and come back.”
But her broader optimism came through in her closing thoughts. “The potential of a post-labor world is tremendous,” she said. “When machines can do the work, humans can focus on what makes us human—compassion, kindness, creativity. That’s the future I want to build.”
