avalw news
William BensonWilliam BensonVIEW PROFILE →

Nobody Told You Retail Changed: Inside the Quiet Rise of Agentic Commerce

tech2026-08-24 · 1 min read · 65 reads

You stopped scrolling. You started asking. And somewhere in between, an algorithm started doing your negotiating, your comparing, and increasingly, your buying.

You stopped scrolling. You started asking. And somewhere in between, an algorithm started doing your negotiating, your comparing, and increasingly, your buying.

Think back to the last thing you bought online. Did you type a product name into a search bar, scroll past a wall of sponsored listings, and click through five nearly identical tabs before deciding? Or did you just ask something — a chatbot, an app, a voice assistant — to find it, compare it, and maybe even buy it for you? If it was the second one, you've already stepped into a version of retail that barely existed two years ago, and most of the industry is still scrambling to catch up with you. The Death of the Search Bar Is Not a Metaphor "Agentic commerce" is the industry's slightly clunky name for a genuinely simple idea: instead of a person browsing a store, an AI agent does the discovering, comparing, and sometimes the buying, acting on that person's behalf. It's a meaningful step beyond the chatbots retailers bolted onto their websites a few years back. A basic chatbot answers a scripted question. An AI shopping agent interprets what someone actually needs, weighs options against price, reviews, and fit, and takes real action toward completing the purchase — sometimes without the shopper ever opening a retailer's website at all. The shift happened faster than most predictions expected. AI-driven orders grew roughly fifteenfold between January 2025 and January 2026, and AI-referred traffic to US retail sites jumped more than 800 percent year-over-year around the most recent Black Friday. Nearly two-thirds of consumers now say they plan to use an AI chatbot for shopping this year, with close to a quarter of them saying it's already their default way to shop — not an experiment, a habit.

How This Actually Works Right Now, Not in Theory

Every major platform has taken a slightly different route to the same destination. OpenAI briefly built full checkout directly into ChatGPT conversations, letting people complete a purchase without ever leaving the chat — but by early 2026 it had pulled back from that approach, shifting focus toward product discovery and comparison inside dedicated retailer apps rather than one-click in-chat buying. Walmart's own data help explain why: purchases completed inside a chat window converted at roughly a third the rate of shoppers redirected to Walmart's actual site, suggesting people still want a familiar checkout screen for anything that isn't trivially simple.

Amazon took a more sweeping approach, retiring its earlier Rufus chatbot entirely and merging it with Alexa into a single tool called Alexa for Shopping — a genuine agent that can automate deal-hunting, assemble a cart, and execute routine purchases using a shopper's own year of price history, now available to every Amazon customer without requiring a Prime membership. Google, for its part, launched something called the Universal Commerce Protocol at a major retail conference in January, an open standard meant to let AI agents from any company browse, compare, and buy across a huge range of retailers rather than staying locked inside one company's walled garden.

Groceries Turned Out to Be the Real Battleground

Fashion and electronics were the obvious early targets for shopping AI, since a jacket or a laptop doesn't change from hour to hour. Groceries are a much harder problem — inventory turns over constantly, substitutions are inevitable, and a wrong call about produce or a missed allergy note actually matters. That difficulty is exactly why grocery has become one of the most closely watched fronts in the whole shift.

Instacart became the first grocery partner to plug directly into ChatGPT's shopping tools, letting people order groceries through a conversation instead of an app. DoorDash pushed its own AI shopping integration well beyond food, expanding into nearly every category of local commerce. Albertsons, meanwhile, has been building a proprietary AI shopping assistant with a genuinely bold internal target: cutting the average time it takes a customer to complete a weekly grocery run from roughly 46 minutes down to about 4.

The Retailers Racing Not to Be Left Behind

What's notable is how quickly traditional retailers, not just tech platforms, have signed on rather than fighting the shift. Walmart, Target, Wayfair, Ashley Furniture, and Home Depot all became early partners on Google's open commerce protocol, and Best Buy joined as a Google AI Mode partner specifically. Macy's endorsed the same open standard at launch. Walmart has gone as far as training over two million of its own store associates on how AI shopping tools now factor into the customer journey, treating it as a genuine operational shift rather than a marketing experiment.

The pattern across nearly every major retailer right now is the same: get a product's data structured and visible to AI agents immediately, because a listing an AI agent can't understand is, for a growing share of shoppers, a listing that effectively doesn't exist.

A growing share of the
boxes arriving on doorsteps in 2026 were never clicked "buy" on by a
human being at all — an AI agent placed the order.
A growing share of the boxes arriving on doorsteps in 2026 were never clicked "buy" on by a human being at all — an AI agent placed the order.

Who Actually Wins When an Algorithm Does the Shopping

There's a real tension sitting underneath all this convenience. When an AI agent picks a product on someone's behalf, the brand that gets chosen doesn't necessarily win a loyal customer — the platform running the AI agent does. A shopper asking for "a good moisturizer under $40" never sees a brand's carefully designed storefront or hears its pitch; they see whatever the agent decided was the best match, based on structured data most consumers never think about. Products with fully completed technical markup are, by one industry estimate, over six times more likely to get picked by an AI agent than products without it — meaning invisibility to an algorithm is quickly becoming as damaging as being hard to find on a shelf once was.

There's a cost question too. When OpenAI briefly ran in-chat checkout, it charged merchants a 4 percent transaction fee on top of standard payment processing — meaning a $100 order could carry more than $7 in combined platform and processing costs before the retailer sees a cent. Multiply that across the trillions of dollars analysts expect this channel to handle by 2030, and it's easy to see why so many retailers are racing to shape these new rules early rather than have them shaped without their input.

A Few Honest Questions People Are Actually Asking

Is an AI shopping agent actually going to get me a better price? Often, yes, for straightforward comparisons — that's genuinely what these tools are built for. For nuanced purchases (fit, quality, personal taste) it's still worth double-checking a recommendation rather than trusting it blindly.

Can an AI agent really buy something without me confirming it first? Most current tools, including Amazon's and Google's, still require some form of confirmation or pre-set spending rules before completing a purchase, though the trend is clearly moving toward more autonomous "set it and forget it" buying for routine, low-risk purchases like restocking household basics.

Will smaller brands get squeezed out entirely? Not necessarily out, but they will need to adapt — brands with clean, structured product data are already outperforming ones without it in AI-driven recommendations, regardless of size.

What This Actually Means for How You Shop

The honest takeaway isn't that browsing is dead — plenty of people still enjoy scrolling through options, especially for anything they care about aesthetically. What's genuinely changing is which purchases people bother browsing for at all. Routine, repeat, low-stakes buys — the household staples, the same brand of coffee, the predictable grocery restock — are exactly the category AI agents are absorbing fastest, because there's little upside to spending human attention comparing products someone already knows they want.

That leaves actual human browsing time concentrated on the purchases that genuinely benefit from it: the things people want to see, touch, or feel excited about before buying. In a strange way, letting an algorithm handle the boring 80 percent of shopping might end up making the remaining 20 percent more enjoyable, not less.

A note on the numbers

Adoption and revenue figures in this piece are drawn from multiple 2026 industry trackers and company disclosures, which update frequently as platforms adjust their approach — as several already have this year. Treat the exact figures as a snapshot of where things stood in August 2026.

William Benson
Stay updated
William Benson
Subscribe to get an email whenever William Benson publishes a new story. No spam, unsubscribe anytime.
William Benson
WRITTEN BY THE AUTHOR
William Benson
2026-08-24 · 1 min read · 65 reads
View profile →
VERIFY THIS STORY
ASK AI
MORE FROM William Benson
Report this articlesupport@avalw.com