Everything is already on a planagram. It would be trivial to do.

Being generous, I could see A use case for translating whatever the customer says (because how often have you known something exists, but not what it's called?) into an actual product and then looking it up in a proper database. This, though, is bound to fail.

I don't know if it exists in the UK, but there's a craft store chain in the US called Michael's, and when Covid kicked up they changed a lot about how their business is done.

So many things went online-only after 2020. I'm a life-long artist/crafter, and let me tell you, the people making decisions at this place clearly have no idea how crafters' minds work. We don't always go into stores with a particular goal in mind - sometimes we go in just to see what's there and get inspired to try something new. We see things we didn't think about, but once we see it, we know exactly what we want to make/how we want to work an item into a project.

But in order to do that, things have to physically be in the store. We aren't going to impulse-buy something that isn't there. Obviously.

As far as art supplies go, I'd much rather buy something I can physically interact with before purchasing - this tool says it has an ergonomic grip, but does it actually feel better in my hand or is it just hype? Can I trust a screen's color rendering to faithfully represent the hue of this product? Will the feeling of this yarn be comfortable to wear as clothing?

Sometimes, we have to experience a product to really decide if we want it. Some art supplies might be fine to order online, sight unseen, but to rely on that alone (as a seller) is absolutely foolish.

I guess my wallet should be thanking Michaels, but I'm too frustrated by so many things becoming "online only" to really appreciate the inadvertent savings. I'm far from the only crafter/artist that follows their energy in the moment, and if I go to an arts and crafts store and get told I need to order a product online and wait for it, that can be enough friction to scrap the entire idea. I need to follow my artistic energy when it occurs. If I put it off til later, the project might never happen - because I'll be onto something else by then.

But if the company really insists on misunderstanding their customers and shooting themselves in the foot, I'm not gonna stop them.

1. They could have an computerised lookup of location that isn't AI based very easily. It's just a stock location database.
2. Shops move their stock around to get you to search the store and drive sales of other items " you didn't know you wanted".

My biggest experience of this is that those systems just aren't updated. So they will have rearranged a bunch of stuff and now the website is wrong. When they're doing it right, the same system that generates the printouts to tell store workers where to move things is also connected to the site. But there's still errors. An AI isn't going to solve that in any meaningful way, you just need to pay someone to check the locations of the top 100-200 selling items. Or niche items or whatever works to reduce the number on enquiries. It's pawbably much cheaper to do that once a month, or once every three months, than pay OpenAI for a huge contract. Have that person rotate through all the stores in a region. But that would be too bloody smart for a business to do.

Also, Google turns out seriously shitty answers nowadays, and the sponsored keyword ads are more irrelevant than ever.

I mean.. this. Supermarkets aren't that big. If you're recording where everything is shelved (and that's a big IF, I doubt it's worth it), then a good old search will do the trick just fine.

But also, forget even asking. I would assume supermarkets want you to roam around the store looking for things, because... well, that's how you bump into crap you don't need and buy it anyway. Seems like a weird lose/lose.

I kinda hate that society is being increasingly designed around people being antisocial.

I like it in some ways, but in many it's just freaking sad. Every time I go to my local grocery store, people would rather wait 5 minutes to start self checkout rather than zero wait to have a person check them out. I find it really bizarre. I highly doubt most of those people have the levels of legit social anxiety I have. I don't really have an explanation other than they're used to avoiding people. Maybe COVID has a lot to do with it.

Exactly, this isn't a difficult problem to solve. If it's even a problem in the first place, because a sensible layout with good signage is enough for 99% of customers. A website to navigate a store is probably not even worth it in terms of net time savings, but it's convenient for customers and makes them more likely to go to your store if they can see something is in stock before they even leave their home, which is where the real return on investment is. An LLM can't give you live stock updates, so that's half of the value proposition gone.

Honestly, that is not a bad use of AI. It's close to what I would consider the "ideal" use of the technology.

Yep. Pay staff nights/overtime to do a job that could be done during normal hours. That's a definite recipe for success.

So if something g runs out the customers have to wait until tomorrow for it to be restocked on the shelf? Great idea.