The New Arrivals Page That Answers Sizing Questions at 11 p.m.

The question you keep answering after close

A shipment lands Tuesday. You steam it, tag it, get it on the floor by Thursday. The part where somebody photographs it, writes it up, and posts it? That's a 9:30 p.m. job, and it loses to counting inventory, running payroll, and going home.

Meanwhile the same three questions arrive at all hours. Do you have it in a medium. Does it run small. Is it still there.

Those are website questions, and you are answering them by hand, one text at a time. Every one of them is a sale that either happened or didn't, decided in the ninety seconds before somebody got tired of waiting and opened a different tab.

Something shifted this year that makes that gap more expensive than it used to be. Adobe's analytics, reported in June, found AI-referred traffic to U.S. retail sites grew 138% year over year in May 2026. The more interesting part: that traffic converted 54% better than non-AI sources, and those shoppers spent 53% more time on site and viewed 23% more pages per visit. People arriving from an AI assistant tend to show up already knowing what they want.

They only show up if the assistant can read your site. Adobe's own content visibility scoring put the average U.S. retail product page at 66%, meaning roughly a third of what's on those pages isn't written in a form an AI can actually use. Home pages did better, at 75%.

Google went a step further. For the 2025 holidays it rolled out a feature where Google itself calls local stores to confirm whether an item is in stock, because that information isn't published anywhere findable. Merchants can opt out. But sit with what it implies: the availability answer is valuable enough that Google built a phone workaround for stores that don't post it. Then in January, at NRF in New York, Sundar Pichai announced the Universal Commerce Protocol, an open standard co-designed with Shopify for how AI agents read merchant catalogs and complete checkout, with endorsements from Target, Walmart, Macy's, Visa and Mastercard.

Build one page, correctly: New Arrivals

Don't start with a full site rebuild. Start with the single page a boutique lives on.

The Paglow Website service works from a generated spec first — what you actually carry, how you talk about it, who's walking through your door — then design direction, page-by-page copy, and a live site. For a boutique in Allen County, the page to get right before anything else is New Arrivals, and it has to carry the detail you currently hand out by text message:

• What came in this week, dated, not a gallery left over from last spring
• Sizing in plain language: runs small, model is 5'6" in a medium, actual bust and inseam measurements
• Fabric, stretch, and care — the second question, every single time
• Whether it's on the floor in Lima right now or online only
• How to hold it: text, call, or a form that reaches your phone

Written copy, not just photographs. A carousel of pretty images tells an AI assistant nothing, and it tells a shopper at 11 p.m. barely more.

What this actually changes

Three things, none of them theoretical.

The after-hours questions thin out, because the answer is already published. You'll still get some. You'll stop getting the same one forty times a week.

Posting gets shorter. Once the page exists with a real structure behind it, adding a drop means filling in fields you already know — brand, size run, price, fit note — instead of composing something from nothing at closing time. Ten minutes, not the hour you never had.

And you finally get a list. The signup at the bottom of that page is the first real record most boutiques ever build of who their regulars are. Today that record lives in your head and in a phone full of threads. When a shipment lands that three specific customers would love, a list lets you tell them without scrolling.

None of this turns you into an e-commerce operation. It takes what you already know — how a brand fits, what's on the rack, when the truck comes — and puts it in writing, on a page, where a person or a machine can find it at midnight.