Could you automate an ice cream van?

I was on a walk this morning when I had a weird thought: could you automate the ice cream van business?

Not the driving and not the scooping, but the part that actually loses money.

Here is the real problem with that business. You pull into a neighbourhood, you play the music, and then you hope somebody heard it. That is the whole model, a guy in a truck betting that the right kid is in the right yard at the right moment with a parent who has cash. Everything else about the business is fine, because the product sells itself and the margins on a Rocket Pop are not the issue. What hurts is that demand is invisible until you are already parked, and by then it is too late to go somewhere else.

So you fix the invisible part.

Put a big weatherproof QR sticker on the side of the van that says track my van live and get a text when I'm close. That is the whole pitch, and because it is written on the side of a vehicle that drives past hundreds of houses a day, your acquisition channel is already paid for.

The sticker points at a one-page site, which you can build free on Carrd. It needs two things and nothing else: a live map showing where the van is right now, and a box where a parent puts their phone number and their street. There is no account, no app to download and no login.

The map updates itself because of Glympse, which shares your phone's location into a page while you drive, so you start it when you leave in the morning and forget about it. The van moves, the dot moves, and any parent who bookmarked the page can see you coming without you doing a thing.

The phone numbers go into an SMS tool grouped by neighbourhood, and that grouping is the part that matters. Before you turn into a specific area you tap one button, and everyone on that list gets a text: the van is entering Oak Bay now, see you in five minutes. Five minutes is enough time to find shoes and get to the curb, so you stop hoping and start announcing.

The other half of this is stock, which is the quiet way a truck business bleeds: you run out of the thing everyone wants and keep driving around with a freezer full of the thing nobody wants.

Square's free POS app on an iPad, with a cheap Bluetooth reader, handles it. You load your menu in and enter your starting counts at the beginning of the day, like 50 Rocket Pops and 40 Drumsticks, and every tap on a sale subtracts one. Square already sends a low stock alert on its own, so you are not building that part. When the Rocket Pops drop under ten, a free Zapier rule catches that alert and turns it into a pre-written order email to your supplier, so all you ever do is approve an order that was already written.

There is an AI part too, if you want it. You could vibe code the tracker yourself instead of assembling it from Carrd and Glympse and an SMS tool, which means the map page, the phone broadcasting its location and the alerts firing at the right street all get built rather than bought. Whether that is worth doing depends on how much you like building things, because the off-the-shelf version works either way.

What I keep coming back to is that none of this changes the ice cream: same truck, same freezer, same product, same route. What changes is that you have customers waiting before you park. The old version of the job is drive and hope, and the new version is drive and tell people you are coming.

I have no idea what an ice cream truck actually brings in, and that is the number that would decide whether any of this is worth a weekend.