They told us we could build an online store with free AI — what’s the catch, and what’s missing to actually sell?

We jumped in with the exact same excitement last summer. The AI built the site for free, but then they wanted 24 EUR a month just to connect a custom domain. Plus an extra 1.5 percent cut on every sale via their payment gateway. In the first month, across 40 orders, transaction fees and addon costs came to 110 EUR—the so-called free route ended up way more expensive.

AI isn't doing anything magical here; under the hood, they just slapped a copywriting bot onto an old-school drag-and-drop website builder. The whole point is vendor lock-in. Since you can't export your raw database, whenever you decide to migrate elsewhere down the road, you'll be starting entirely from scratch.

First off, don't rush to burn through that 1.500 EUR ad budget. Buy your own domain and standalone web hosting, then look into open-source e-commerce platforms. Use AI solely for drafting your product copy and cleaning up product visuals instead of trusting it with the entire backend infrastructure.

yeah i watcehd those exact hype videos, they make it seem like packages magically ship themselves once you hit a button lol but tbh i was thrilled when the layout took five minutes, but all my enthusiasm died when i realized i couldnt even hook up our local banks virtual POS.

if u sell in spain there are lssi-ce and tax office rules to follow. tbh relying on privacy policies written by a free ai is begging for fines on the very first consumer complaint, never leave compliance to an ai.

The biggest technical roadblock with these platforms is that you get zero database access. Your customer list, order history, and product catalog stay held hostage on their servers. When you eventually need to hook it up to local bookkeeping software via an API, they will force you into an expensive enterprise tier.

The 4 concrete costs you will face before making a single sale are: 1) Custom domain and SSL setup fees, 2) Payment gateway processing commissions, 3) EU-compliant legal and cookie consent policies, 4) Invoicing software capable of booking retail transactions properly.

quick question though, if we use the AI-generated stock photos it creates for the storefront, will we run into cpoyright issues down the line?

edit: I wrote something wrong above, sorry about that.

I don't think this advice fits everyone. Every "we'll look at it later" in a meeting means unbilled work.

Of course, it varies if your situation is different.

Quick summary for newcomers: Solutions that work at a small scale collapse when you grow; I learned this late.

That's all, sorry if I went on too long.

we need to make a distinction here then when making a decision first look at what data you have on hand.

that's all, sorry if I went on too long.

Thanks for writing this, that's the right way. The cheapest quote is usually the least thought-out one.

That's all, sorry if I went on too long.

I was thinking the same thing. Payment schedules should be tied to project phases, not calendar dates.

If you have questions, write them; I'll answer as best I can.

My question might sound amateurish sorry about that. Taking notes for two weeks yields better results than a six-month estimate.

Correct me if I'm wrong.

Thanks for writing this, that's the right way. Processes without records never improve, because you don't know what to fix.

When we decide without measuring, we always end up in the same place. Proven by experience.

I didn't know that.

i have a question. weekly written progress reports are much more useful than asking for dates.

if you get three different answers on a topic the question was asked wrong. honestly proven by experience.

We need to take it step by step. Most time waste accumulates in tasks waiting for approval.

Ask who prepared the quote and who will actually do the work. Hope this helps.