The introduction of AI agents like ChatGPT and Perplexity into the shopping journey is more than a tech headline — it’s a structural shift in how demand is created, influenced and converted. These tools are starting to compress the traditional funnel, blurring the line between demand generation and demand capture in ways retailers and brands need to plan for now.
From Curiosity to Cart in One Prompt
Historically, demand generation efforts sparked interest through upper-funnel tactics like display media, influencer campaigns, or social discovery, while retail media and search captured demand closer to purchase. But AI agents merge those stages.
With a single query like “What’s the best laundry detergent for sensitive skin?”, shoppers are served a shortlist of products with pricing, sentiment, and direct links. That’s top, middle, and bottom funnel collapsed into one interaction — and it’s guided by AI, not ads.
Retail Media Must Evolve to Influence the Agent
Retailers and RMNs are uniquely positioned for this shift. They already have what AI agents need — structured product data, behavioral signals, and conversion-focused media tools. But those assets must now be deployed with the speed and intelligence AI ecosystems demand.
Retailers and RMNs already have strong tools for demand capture: intent signals, high-quality product data, and robust targeting platforms. But AI agents aren’t simply looking for paid placements — they’re crawling content, reviews, and structured data to decide what’s most relevant.
To stay influential, retailers need to:
- Ensure product detail pages (PDPs) answer the kinds of questions AI agents might receive
- Optimize back-end attributes, not just front-end creative
- Shift from static content libraries to dynamic, AI-ready content pipelines
This is how demand generation can be influenced in an AI-first environment — not by visibility alone, but by data completeness, clarity, and consistency.
Retail media’s role isn’t fading — it’s evolving. AI isn’t replacing it; it’s demanding more from it.
Who Owns the First Click Now?
This shift also introduces uncertainty:
- What determines which products are recommended?
- How do AI agents evaluate across retailers?
- Will there be pay-to-play models that brands can influence, or only earned relevance?
Perplexity’s Merchant Program suggests a path forward for those looking to participate directly, but most retailers and brands will need to adapt with zero-click discoverability in mind. The path to purchase is now more agentic and less human — and that means less time to influence, but more pressure to be found.
What This Means for Marketers
The rise of AI agents doesn’t eliminate demand generation or capture — it compresses them. For those in retail, CPG, or ecommerce, the goal now is to optimize for decision-speed environments where your product may be recommended, rejected, or replaced without a single click on your media.
This is where retail media has the chance to evolve from placement-driven to performance-optimized. From static to dynamic. From slow-moving to real-time.
Action Steps:
- Align media and content strategies around the types of questions AI agents are fielding
- Prioritize speed, accuracy, and depth in PDPs and product feeds
- Partner with retail media teams to explore how AI readiness can be layered into content ops
- Begin testing how your products surface across different AI tools (e.g., ChatGPT, Perplexity)
Final Thought
AI agents won’t destroy the funnel — they’ll redefine who owns each moment. The question isn’t whether AI changes the journey. It’s whether the infrastructure behind it can keep up. The brands and retailers who can build visibility within these new environments — and better leverage their data and AI tools across existing retail media networks — will be the ones who win in both demand generation and demand capture.
Brandon Viveiros is chief commerce officer for AdFury.ai and has more than 15 years experience in digital and shopper marketing. He is a champion of innovative learning and empathetic leadership. A version of this commentary originated on LinkedIn.
