$8k in sales, two weeks after launch.

Poolboy makes three pool floats. A lounger for lying flat, a round float you sit in, and a floating hammock. All of them fabric, none of them sticky.

It's an e-commerce brand we're partners in, which makes it the cleanest test we have. Our money, our brand, and the same system we build for clients.

Every product shot and every pool scene on the store was made with AI. Every colourway, every angle, the same studio background behind every hero shot, so the range looks like one brand rather than three suppliers.

Colour picked from swatches, not a dropdown. Bundle tiers backed by real automatic discounts, two floats or the full set added to the cart in one tap. Google reviews under the title, free shipping nationwide, the returns and warranty line where people look for it, and an add to cart bar that follows you on a phone.

Under that, the spec, the dimensions and a Before you buy section that handles the objections a salesperson would.

One Meta campaign, one ad set per product, an ad per colourway. 128 different ads went through it.

The point was never to guess the winner. It was to let the numbers pick it. That formula is now found. From here the job is to scale what works and keep testing the next one.

Google Ads and Merchant Center run alongside, so the floats show up when someone is already searching for one.

The welcome, the saved cart, the checkout reminder and the rest. Sixteen emails, written once, sent every time someone qualifies, without anyone remembering to send them.

AI made the photos. Automation runs the follow-up. The ad testing happens at a volume no small team would do by hand, with the apps doing the repetitive work most people would call agents.

Two weeks in, $8k in sales. And the next brand got the same build in a week.