• by Laurie Sullivan , Staff Writer, Yesterday

Agencies and ad-technology companies are rolling out simulation tools to tell brands how to get the AI-based agents to recommend their products to consumers. 

Jellyfish, an agency under The Brandtech Group, released one on Tuesday called "Agent Shopper." The simulation tool analyzes AI-driven shopping behavior across search, social, and retail platforms. It demonstrates to brands how agentic technology researches, compares, and recommends products before providing that feedback to consumers. 

Nick Emery, CEO at Jellyfish, believes the 2025 holiday season will be the last in which people alone make purchase decisions. By next year, he said, AI agents will research and reason what drives commerce.

Emery told MediaPost via email that the last holiday season for human only decisioning really began in 2024. 

"Consumers consult LLMs for support on shopping decisions over 4 billion times each month," he wrote. "We’ll see a flip next year where agent-assisted and agent-completed purchases outpace human-based transactions."

Digital ads, search, and other content still matter, he wrote. LLMs observe the digital content, including ads, search, webpages, social channels.

Emery views LLMs as a "new crucial and increasingly important audience that content must attract and influence."

Agent Shopper, tied with Jellyfish's LLM insights platform, "Share of Model," identifies how agents use the internet, including advertising, to make recommendations and purchases.

As of July, more than 50 million daily ChatGPT prompts are shopping oriented and more than half of shoppers use AI assistants for shopping tasks.

"This is the largest transfer of marketing power in a generation," Natasha Wallace, global chief solutions officer at Jellyfish, wrote in an email to MediaPost. "Brands can no longer simply buy their way to visibility. AI-mediated customer journeys will determine what brands are recommended and there’s a real sense of urgency to take control of those recommendations."

For brands, the rise of AI agents in commerce will become a fundamental shift that moves competition and ad creative away from traditional marketing and emotional storytelling toward product characteristics and algorithmic visibility.

When AI agents take over purchasing decisions, the product characteristics -- which could include reviews and ratings -- will become critical for a brand's success. AI will prioritize objective, machine-readable data over emotional branding.

Agent Shopper simulates behavior and enables brands to predict this behavior before it occurs, identifying where and how AI systems choose products and helping brands to discover which factors shape AI recommendations and which updates have the greatest impact on visibility and conversion.

Recent Adobe data predicts that AI-driven shopping traffic will increase by 520% year-over-year during the 2025 holiday season.

As AI tools increasingly influence what consumers see and buy, marketers must rethink how they build visibility, trust and relevance across every digital media channel.

Findings from a Jellyfish Agent Shopper study show how and when consumers are shopping for electronics, as brands including Apple, Samsung and Sony compete through agentic services.

The analysis reveals that agents make decisions on behalf of consumers based on transparency, functions and consumer consensus rather than traditional marketing influence.

The path to purchase highlights that “multi-retailer presence, credible peer proof, and structured product data” are becoming new drivers of digital commerce. Reviews, YouTube unboxings, and TikTok demos typically outscore official brand public relations.

Today’s agentic path to purchase is a contest of transparency, innovation and peer-reviewed trust, where product attributes must be verifiable and consumer sentiment must be real, not manufactured.

Media and review sites fill the gap on expert testing and comparative insights, alongside retail marketplaces.

Joe Mandese, MediaPost editor-in-chief, wrote about Interpublic's simulation studio that launched last summer, for example. He explained how the tool is powered by agents-based research capable of simulating human responses to any scenario, including ad campaigns at a "speed and accuracy unimaginable by human-based research methods."

It is based on technology, software and research methods developed by AI startup Aaru, which uses multi-agent AI systems to create simulations of human behavior for predictive intelligence.

Brand-specific AI chatbots ready to help consumers shop include Amazon Rufus, Sephora Virtual Artist, Walmart Shopping Assistant, Lowe's Mylow, and many more from companies including H&M, Levi's, Tommy Hilfiger and Lidl.