AI Discovery Is Becoming a Revenue Channel. Retailers Should Treat It Like One.
Two reports published in October caught my attention.
Google found that consumers using AI search visit 12 touchpoints during their shopping journey, nearly three times as many as those who don’t use AI. They’re exploring more options, asking more specific questions, and making decisions with greater confidence.
Adobe reported that AI-referred traffic to U.S. retail websites grew 71% year over year in September. Those visitors converted 90% better than non-AI traffic and generated 83% more revenue per visit.
A little over a year ago, AI-referred visitors were converting worse than other visitors.
The numbers tell us something important about how shopping is changing.
AI isn’t necessarily making the shopping journey shorter. In many cases, it’s helping consumers explore more possibilities, compare alternatives, and make better-informed decisions.
And when those shoppers arrive at a retailer’s website, they’re increasingly ready to buy.
For retailers, I think this changes the conversation. AI discovery is becoming a revenue channel, and we should start treating it like one.
From visibility to revenue
Over the past year, retailers have understandably focused on AI visibility.
Are our products showing up in ChatGPT? Is AI recommending our competitors? How do we improve our citations and rankings?
These are useful questions. Retailers need to understand how AI systems see their products and brands.
But knowing where you stand is only the beginning.
A retailer can have excellent visibility and still lose the sale. Or identify hundreds of opportunities without having the resources to act on them.
The real challenge is understanding where customer demand is emerging, deciding which opportunities are worth pursuing, and doing the work to capture that demand.
Then measuring whether it actually generated revenue.
We’re seeing this firsthand with retailers at Spangle.
What we learned with Express
Express was already using AI visibility tools. The team had identified more than 800 recommendations for improving how its products appeared in AI-driven discovery.
The problem wasn’t a lack of information. It was turning that information into action.
Like most retailers, Express has a business to run. Asking its teams to work through hundreds of recommendations isn’t a scalable solution.
We took a different approach.
Spangle’s AI agents, trained on the Express brand and catalog, identified customer demand opportunities and recommended actions. The Express team decided what to pursue and approved what went live. Our agents then created shopping experiences designed around those opportunities.
In September, Express saw AI-attributed revenue grow 14.5x compared with the preceding 30 days. AI-referred sessions increased 6.9x, and the conversion rate for those sessions grew 6.7x.
These are early results from a before-and-after comparison, not a controlled measure of incremental lift. But they show what becomes possible when a retailer moves beyond measuring visibility and starts acting on demand.
The lesson for us was straightforward. Retailers don’t need another long list of things they should do. They need the ability to identify the opportunities that matter and act on them.
A new channel needs a different way of operating
Retailers have spent decades building sophisticated capabilities around paid search, social, email, and other acquisition channels.
They know how to allocate budgets, measure returns, and improve performance.
AI discovery deserves the same commercial discipline. But I don’t think we can simply apply the existing playbook.
Traditional search largely organizes information around keywords and links. AI increasingly understands what a shopper is trying to accomplish, evaluates alternatives, and recommends specific products.
Consider someone asking an AI assistant for an outfit that works for a business trip, or a pair of shoes comfortable enough for a day of walking but appropriate for dinner.
The opportunity isn’t necessarily to rank for a particular keyword. It’s to understand that shopping need, determine which products are a good fit, and compete for the recommendation.
Google’s research suggests consumers are exploring more possibilities through AI, not fewer. That creates more opportunities for retailers to be considered, but also more shopping situations they need to understand and serve.
Doing this across thousands of products and countless shopping situations is difficult with the systems and processes retailers use today.
I believe AI can fundamentally change how this work gets done.
Instead of separate teams and systems identifying opportunities, creating recommendations, implementing changes, and analyzing results, intelligent agents can connect those activities.
They can understand demand, recommend actions, execute what the retailer approves, and learn from the outcomes.
That’s the approach we’re building at Spangle.
Retailers remain responsible for their business decisions. AI can do much more of the work required to execute them.
And the shopping journey is changing again
This week, Amazon announced that select customers can now text Alexa for Shopping to handle shopping requests.
Customers can ask Alexa to find products, check prices, add items to their carts, and complete purchases.
It’s an interesting development because it moves AI beyond helping customers research products. The agent can now take action on their behalf.
I expect we’ll see more of this.
Today, much of the measurable AI discovery opportunity comes from shoppers researching products through AI and then visiting a retailer to purchase.
As personal shopping agents become more capable, more of the comparison, selection, and purchasing process may happen outside the retailer’s website.
That raises a harder question for retailers.
What happens when you no longer control where the customer starts shopping, which products they consider, or even where the transaction happens?
Retailers won’t control every part of that journey. But they can control how well they understand their products, customers, inventory, pricing, and the demand they’re positioned to serve.
They can build intelligence that represents their business and helps them compete, regardless of where the shopping journey happens.
I don’t think retailers should wait for the agentic commerce ecosystem to settle before getting started.
The capabilities needed to compete for AI-driven demand today will become even more important as personal agents take on a larger role.
The opportunity is here
AI-referred traffic is still a relatively small part of overall retail traffic. But its growth and improving economics deserve attention.
Google’s research shows how AI is changing the way consumers explore products. Adobe’s data shows that AI-referred shoppers are becoming increasingly valuable.
Our experience with Express suggests that retailers can do more than measure their presence in this new channel. They can identify opportunities, take action, and generate measurable business results.
For retail leaders, I’d start asking the same questions we ask about any other revenue channel.
Where is the demand? Where can we win? What are we doing to capture it? And how much revenue are we generating?
AI discovery is becoming a revenue channel. It’s time to build the capabilities to compete in it.