Next week, I'm releasing a three-part series with Anne Hallock from Mirakl Ads on a theme I don't think gets enough airtime: retail media as a capability that pulls forward technology investments the business needed to make anyway. Catalog quality. Customer data infrastructure. Identity. Measurement. None of these problems were invented by retail media — but retail media is increasingly becoming the lever that finally gets them funded.
That conversation was still ringing in my ears when I came across a Total Retail Talks podcast interview that resurfaced in my feed this week. The episode itself aired in February, but the framing felt fresher than ever. Because the same dynamic that Anne and I had been talking about — retail media pulling forward overdue work — is now happening with AI-enabled commerce.
The interview was with Nick Lezin, SVP of eCommerce at Ashley Furniture, North America's largest furniture brand. He walked through why Ashley decided to launch a fully transactional shopping experience inside Perplexity late last year, but more interestingly: what it's forcing them to fix internally.
The data justified the risk before the hype did
The standard objection to early AI commerce integrations goes something like: the traffic is too small to bother with, the technology is unproven, and you risk training customers to shop in places where you can't influence the experience.
Nick's response is grounded in what he can actually measure. Ashley has been watching referral traffic from AI sources closely, and at the time of recording, it was on an exponential curve from the start of 2025 with no sign of slowing down. But — and this is the part I appreciated — he was honest that the absolute volume is still small.
This is why he frames Ashley's move as low risk, high reward rather than a panic pivot. You're not betting the company. You're buying optionality on a curve that's pointing up.
One caveat to Nick's framing, which I've written about before: AI-influenced traffic is almost certainly being undercounted by most retailers' analytics, because referral headers get stripped when shoppers click through from apps like ChatGPT and Gemini. So the exponential curve Nick is seeing might be steeper than it initially looked.
Curious about the live experience, a few months in, I tried it out. I ran several furniture queries and didn't get an Instant Buy-eligible Ashley product to surface. But I want to be careful not to overinterpret that. Perplexity's Instant Buy checkout is conditional: only available for select merchants on eligible items, in eligible regions, and on the right kind of query. When those criteria aren't met, Perplexity hands you off to the merchant's website to complete checkout. So my not finding an Ashley Instant Buy says more about what early agentic commerce still feels like today — patchy and conditional.
The real story is what it's forcing Ashley to fix
But the real innovation for Ashley isn't coming from a trickle of AI-native sales. It's how this whole movement is a catalyst for improving multiple touchpoints of the company's current ecommerce and store experience.
"This is an area where introspectively we feel we have a ton of opportunity for improvement," he said, talking about how often Ashley's product imagery, descriptions, and metadata get refreshed. "We've probably been overly static to date... for products that may stay online for three to five years."
Retailers know that a marketplace model can dramatically boost product assortment, shopper engagement, and total revenue. But, to get the most out of your marketplace, you need an ad tech solution that can really engage sellers. Mirakl Ads is powering the future of retail media for leading retailers — to activate both 3P sellers and 1P brands.
He's describing a catalog "thousands and thousands of items deep" where the content has, by his own admission, been allowed to gather dust.
This is the work nobody wants to fund. It doesn't immediately make money the way a retail media program does. It's not urgent in any single quarter. It's a hot potato that gets passed between merchandising, marketing, and ecommerce teams for years. Brands often take multiple months to update a single PDP once you factor in translation workflows, legal review, and brand verification.
But now Ashley has a reason to do it. The Perplexity integration only works if the underlying product data is good enough for an LLM to surface and recommend Ashley's products in the first place. And Ashley has publicly committed to re-platforming ashleyfurniture.com in 2026 to be more AI-agent-friendly, alongside expanding internal AI talent and rolling out AI tooling for store associates.
It's a tack that other retailer and brands can take if they know they're behind: it's the gift of the forcing function.
The political gift of an AI mandate
AI-enabled commerce gives ecommerce leaders a reason to ask for budget and attention for things that have needed fixing for years: Product data hygiene. Inventory feed accuracy. The unsexy infrastructure work that has sat at the bottom of every prioritization exercise because it's hard to model an ROI on it.
Tie any of that work to AI readiness and suddenly exciting, urgent, a board-level conversation. The same project that couldn't get funded as "improve PDP quality" gets approved as "prepare our catalog for agentic commerce."
I don't think that's cynical. I think it's how change actually happens inside large retailers and brands. The catalyst can sometimes be a credible external threat that gives internal champions permission to push through work that's been stuck.
Don't look a gift horse in the mouth.
Ashley is that they didn't wait for the tech stack to be perfect before jumping in. They picked one AI partner and started learning. The re-platforming and the catalog work happen in parallel, with the Perplexity integration providing both motivation and a real-world testing ground.
It wasn't a clean sequence of "fix the foundations, then plug into AI." More like "make a small commercial bet that creates internal urgency to fix the foundations."
And it brings us back to where I started. Retail media did this for the last five years — pulling forward catalog, content, and data infrastructure investments that should have happened years earlier. AI-enabled commerce is doing it again now, only on a compressed timeline. Both are, in their way, gifts to the people inside retailers who have been trying to fix this stuff for years and finally have the air cover to do it.
Link to the full Total Retail Talks episode.
