Zero to E-commerce Site in One Day with Specialized AI

I've written before about running a personal AI agent day to day. This post is a case study: what happened when we pointed a purpose-built agent at one concrete goal - take my wife's Etsy shop and turn it into a real, standalone e-commerce site.

My wife runs Simple Page Studios, a stationery and sticker shop that started on Etsy. Etsy is a great launchpad, but every Etsy seller eventually feels the same squeeze: fees on every sale, no control over the customer relationship, and a storefront that looks like everyone else's. The obvious answer is your own site. The obvious problem is that "just build a WooCommerce store" is weeks of work for most small shops - which is why so many never do it.

We did it in a day. The store at simplepagestudios.com went from nothing to a live WooCommerce catalog with hundreds of products - a full Etsy to WooCommerce migration - and the way we got there is the interesting part.

Meet Scooter: a specialized agent, not a general one

The agent that built the site isn't the same one that runs my own business operations. It's a dedicated instance - we call it Scooter - set up specifically for Simple Page Studios, with its own memory, its own saved skills, and its own chat where my wife can talk to it directly.

That specialization matters more than it sounds. Scooter has already learned the shop's house rules for every listing: the exact way product titles are written, which sticker sizes are offered and what each size costs, the file specs the printer requires for die-cut artwork, and even the order product photos should appear in. When my wife asks for a new product listing, she doesn't re-explain any of that - she sends a photo and a few details, and the listing comes back already following house style. A general-purpose chatbot starts from zero every conversation. A specialized agent starts from everything it has already learned about this one business.

How to import products from Etsy to WooCommerce (the way we did it)

The biggest single job was the catalog: the actual import from Etsy to WooCommerce. Rebuilding hundreds of Etsy listings by hand - titles, descriptions, prices, tags, size variations, and every product photo - is the wall most sellers hit. Here's how the agent handled it:

• Mapping first, importing second. Before touching anything, Scooter wrote a field-by-field mapping document: every column in Etsy's listing export (the "Currently for Sale Listings" CSV) matched to a field in WooCommerce's built-in product CSV importer, with explicit rules for the tricky parts - currencies, variations, out-of-stock items, digital products.
• Stable identifiers, no duplicates. Every imported product got a deterministic SKU derived from its Etsy listing, so the import could be safely re-run without ever creating duplicate products. Boring, and exactly the kind of detail that separates a clean migration from a mess.
• Images came along for the ride. First photo as the featured image, the rest as the product gallery, with failures logged per product rather than silently skipped.
• Test batch before the full catalog. A small subset imported and reviewed first; only after that checked out did the full catalog go in.

That is the honest recipe if you want to import Etsy products to WooCommerce without duplicates, broken images, or mangled variations: map first, key every product to its Etsy listing, and test small before you import the whole shop.

The storefront itself is deliberately un-exotic: WordPress, WooCommerce, and a fast lightweight theme, with the brand's colors, menus, and homepage sections layered on top. No proprietary site builder, no monthly platform ransom - she owns the whole stack.

Polish is where the AI earns its keep

Launch day gets the headline, but what has impressed me most is everything since. A store is never "done," and the ongoing pattern looks like this: my wife notices something on her phone, mentions it in the chat - often with just a screenshot - and Scooter ships a scoped, verified fix, usually within minutes. A few real examples:

• A custom mobile menu. The theme's stock mobile menu was a plain list of links. She got three mocked-up options to compare, picked the floating dropdown card, and that's what shipped - with the search box moved inside it and tap-outside-to-close behavior.
• Cart fixes a human tester would miss. A remove-from-cart button that was invisible (white icon on a white background), an empty cart page that looked broken on mobile, a cart icon that went dead during checkout - each diagnosed to the actual root cause and fixed with a few lines of targeted code, then verified live on a phone-sized screen.
• Product page upgrades. Sticker listings got a benefits checklist - waterproof, high-quality vinyl, free shipping, size options - added automatically across all 180 sticker products, including any added in the future, without touching each product by hand.
• A proper contact pipeline. About and Contact pages, a styled contact form that emails straight to the shop's inbox, and menu placement - requested in chat, live the same day.

Beyond the website, the same agent handles production work for the shop: batch-processing sticker artwork to the printer's die-cut spec and generating consistent product mockups at the right dimensions. The website wasn't a one-off project; it's one surface of an operator that supports the whole business.

The safety rails

Same rules I use everywhere and have written about before: the agent drafts, humans approve. New products go up as drafts for my wife's review before publishing. Nothing customer-facing ships without a yes. And after every change, the agent verifies its own work against the live site - checking the actual rendered page, on desktop and mobile, before calling anything done. Trust is earned by verification, not vibes.

What "one day" really means

To be precise: the site - platform, theme, brand styling, and the migrated catalog - went live in a single day. The refinements above happened afterward, in sessions that each took minutes, not evenings. That's the honest shape of AI-assisted work right now: the big lift compresses from weeks to a day, and the long tail of polish compresses from "someday" to "same afternoon."

For a small shop, that changes the math entirely. The reason most Etsy sellers never build their own site isn't the hosting bill - it's the weeks of skilled work. Remove that, and owning your own store stops being a someday project.

If you sell on Etsy, Amazon, or anywhere else rented, and you've been putting off a store of your own - this is exactly the kind of project we do. Tell me about your shop in the form above, or get in touch, and I'll give you an honest read on what it would take.

Dan Kennedy - Owner, Summit Web Consultants