By Jack Wagner, CEO, Hawk Digital
The Shopping Journey Starts Earlier
In our issue on the collapse of publisher traffic, we looked at what happens when people get an answer without visiting a website. Retailers face a related question: what happens when a shopper chooses a product before reaching your store?
Picture someone shopping for headphones for a long flight. They ask an AI assistant to compare comfort, battery life and price. By the time they visit a retailer, they may already have a model in mind. Your site still matters, but some of the work of winning that customer happened elsewhere.
A yStats research summary reports AI use of roughly 62% for product comparison, 23% at checkout and 19% after purchase. Those figures describe different uses of AI, not a conversion funnel. For retailers, the useful question is what happens between a recommendation and a completed order. Product information, available stock and payment all need to work together.
Three Places to Start
Start with the parts of the business that already cause friction for shoppers. AI makes accurate information and a dependable checkout even more important.
1. Make your product information dependable
An assistant comparing products needs usable descriptions, current prices, availability and clear specifications. A missing size or an outdated stock count can undermine the recommendation. Check what your product feed says against what a customer can actually buy. The same yStats summary reports AI referral growth of up to 4,700% year over year, from a low base. That’s a reason to test the channel, not to write off your existing ones.
2. Walk through the handoff to checkout
A recommendation only gets you so far. If the customer has to find the product again, create an account, or wait until the last screen to see shipping costs, you’ve added work at the worst possible moment. Test the full journey: the selected item, its price, available stock, delivery estimate, payment authorization and return terms.
3. Let the shopper approve the purchase
RTB House research, reported by Retail Dive, found that 35% of respondents wanted a human review before an agent bought on their behalf, rising to 44% among baby boomers. Among US millennials, 42% would allow purchases up to $250 if returns were available within seven days. Give people a clear confirmation step and an easy way to correct a mistake. A fast checkout is only useful if the customer understands what they’re buying.
Getting Ready for Agent-Led Shopping
- Check your product feed for missing details and outdated prices
- Test authorized payments and remove unnecessary account-creation steps
- Require shopper confirmation for higher-value or non-returnable purchases
- Update inventory often enough to support the promise you’re making
- Choose the agent integrations and standards you can support reliably
Where Hawk’s Enterprise AI Command Center Fits
There’s room for both third-party discovery and a direct relationship with the shopper. Someone may find you through a general-purpose assistant, then need help from your team with fit, delivery or a return. That handoff deserves as much attention as the first recommendation.
Hawk’s Enterprise AI Command Center is being built around that continuity: connecting customer conversations and giving teams a clearer view of the journey. For a retailer, the starting point is practical. Decide what catalog and inventory information an assistant needs, what it may do, and when a person should take over. Build around a specific customer problem, then test whether the experience helps people complete their purchase.
Final Thought
The lesson carries over from our ad operations issue: the experience is only as reliable as the systems behind it. A helpful recommendation can’t make up for an unavailable product or a checkout that loses the customer’s selection.
Forecasts collected by yStats put the global opportunity at $3 trillion to $5 trillion by 2030, and McKinsey and ICSC project a US market near $1 trillion. Those are estimates, not booked revenue. You don’t need to bet on the headline number to make your product data, checkout and customer support work better with automation.
Start with one product category. Follow a recommendation through to the order, and fix the first thing that gets in the customer’s way.
What We’re Watching
- AI in advertising: The Trade Desk’s Zuma update adds AI tools for audience creation and campaign workflows in Kokai, including access to retail media and ecommerce data. We’re watching how those tools change the work buyers still need to review. (AdExchanger)
- Shoppable CTV: Broadpeak’s Click2Cart lets viewers add a product to an online cart with their TV remote during a streaming ad. The next test is whether that simpler handoff leads to completed purchases. (TV Tech)
- Streaming and ad ops: Ad-supported TV accounted for 71.5% of viewing in Nielsen’s Q2 2026 Ad Supported Gauge, with streaming at 48.2% of that total. Nielsen also flags an upcoming methodology change, which matters when comparing results over time. (Nielsen)
Work With Hawk Digital
Hawk Digital helps publishers, platforms, brands and agencies connect revenue strategy with the operational work needed to deliver it, across CTV, retail media, commerce, ad operations and AI workflows.
Book a quick call with our team →
This article first appeared in Hawk Talk: “Voices of Innovation,” our LinkedIn newsletter. Read the original on LinkedIn and subscribe for new issues.
