Why Your Future Top Customer May Never Actually Visit Your Homepage

For years, retailers have optimized for a single moment: the website visit. How do we drive more traffic to our site? How do we increase time on page? How do we move shoppers from homepage to product detail to checkout? That model, however, is breaking down.

Right now, a meaningful and rapidly growing segment of your highest-value customers are making purchase decisions before they ever reach your site. They're doing so via an artificial intelligence assistant. These AI agents include tools like ChatGPT, Perplexity, Google AI Overviews, and emerging shopping assistants that help consumers research, compare and select products before visiting a retailer’s site. By the time a consumer arrives at your product page, the consideration work is done. They're there to buy.

The implication is simple but uncomfortable: your future best customer may never meaningfully “visit” your homepage at all. That's a fundamental change to the path to purchase. And most retailers are measuring it with dashboards built for a world that no longer exists.

The Shift That's Already Happened

Generative AI-driven traffic to U.S. retail sites increased 4,700 percent year-over-year as of July 2025, the first period when AI-referred traffic was substantial enough to serve as a meaningful baseline. During the 2025 holiday season alone, AI-powered chat services and browsers drove 693 percent more traffic to retailer websites than the prior year. This demonstrates a structural change in how discovery works.

More importantly, the traffic quality is exceptional. Adobe found that AI-referred visitors show 32 percent longer sessions, 10 percent more page views, and a 27 percent lower bounce rate compared to other channels. While ChatGPT referrals currently represent only a small portion of organic search volume, research tracking 94 e-commerce stores across all of 2025 found that ChatGPT-referred traffic converted 31 percent higher than nonbranded organic search.

The reason isn't mysterious. These shoppers have already done their research inside the AI platform. They've described what they're looking for, seen comparisons, and narrowed their options. In effect, the AI has pre-qualified them. They arrive at your product page with the browse-and-discover phase already behind them, demonstrating that these shoppers arrive highly qualified.

Interestingly, today’s data shows these sessions are still longer and deeper than average. That likely reflects a transitional moment, where AI is accelerating discovery but shoppers are still validating decisions once they arrive. As AI agents take on more of that evaluation upstream, we should expect that behavior to shift toward shorter, faster-converting sessions over time.

Why Traditional Metrics Create the Wrong Signal

The paradox retail leaders need to understand is that when AI-assisted discovery is working — i.e., when shoppers are arriving informed and ready to purchase — your legacy engagement metrics go down.

Over time, as AI takes on more of the discovery and comparison work, we should expect page views to decline, time on site to shrink, and pages per session to compress. Every metric we built to reward shoppers exploring your site looks worse precisely because AI has already done the exploration work upstream. If you're reading those signals as underperformance, you're making the wrong call.

Measurement frameworks need to catch up with behavior. For example, instead of rewarding browse depth, start tracking how quickly a session moves from landing to purchase. Track first-session purchase rate from AI-referred traffic as a distinct key performance indicator, and measure revenue per AI-referred session against your organic baseline. With AI-referred revenue per session already running approximately 10 percent higher than nonbranded organic, treating AI as one undifferentiated traffic bucket is leaving strategic insight on the table.

These metrics reflect intent, not just activity. In an AI-driven environment, intent is the signal that matters most.

The New Competitive Surface

If discovery is happening upstream, then competition is no longer confined to your website.

Recent research suggests that among consumers who use AI for shopping, AI agents and conversational assistants are quickly becoming one of the most influential sources of product recommendations, ranking just behind in-store experiences and ahead of channels like social media, friends and family, and even brand sites and apps. With that understanding, the question is no longer whether AI is shaping your shoppers' consideration set, it's whether your products show up in it.

This is not SEO 2.0. Traditional search engine optimization assumes a human scanning a results page, making choices based on titles, snippets and brand familiarity. Agent optimization assumes a model evaluating structured data such as product attributes, pricing signals, review density, and sentiment and inventory status, then synthesizing a recommendation on behalf of a user who has given it clear criteria. The inputs, the weighting, and the feedback loop are all fundamentally different.

What this means is that retailers need to optimize not just for visibility, but for "recommendability."

What Leading Retailers Are Doing Now

Some of the most forward-thinking retail teams aren’t waiting for this shift to fully materialize, they’re choosing to adapt now.

First, they're auditing product data from the agent's perspective. Ask the question: Would a model confidently recommend this product based on what we've published? This means going beyond basic title and image to evaluate attribute completeness, review quality, pricing clarity, and inventory accuracy. If the data is ambiguous or sparse, an AI assistant will recommend something else.

Next, they're tracking AI-referred traffic as a distinct segment. GA4 and most analytics platforms can capture a portion of AI referral traffic today, particularly from known AI domains, though a meaningful share of AI-influenced visits may still appear as direct or organic traffic when users switch sessions. Retailers that can measure this now will have months of trend data when the channel matures. At the growth rates we're seeing, that won't take long.

Finally, they're rethinking retail media investment. If high-value impressions are increasingly happening inside AI environments that retailers don't own or control, then on-site sponsored placements are no longer the only lever worth pulling. The most sophisticated retailers are starting to ask what it means to have a retail media strategy that extends to the moments of discovery before the shopper arrives.

A New Strategic Reality

The homepage isn’t dying, but its role is changing. In an AI-driven world, the purchase decision is often made before the visit. The homepage simply completes the transaction.

For retailers, that means the focus is less on driving visitors and more on ensuring your products can be recommended by AI agents. Retailers that understand this shift and pivot early will benefit from higher-intent traffic, more efficient conversion, and stronger alignment between media and merchandising.

Those that don’t may find themselves optimizing a front door that fewer and fewer customers ever walk through.

Pat Copeland is general manager of Moloco Commerce Media (MCM), the only AI-native ads engine that enables retailers and marketplaces to activate high-performing retail media networks.