so far the best website at this is instacart to the point that if my preferred brand of instant potatoes isn’t where I’m looking it will direct me to the store that has them…
However usually if they have the item they’re showing me is hit or miss because the stores are responsible for keeping their front up to date and it’s only done every day or two or never with a certain local store …
There was the issue where Kmart or similar was able to figure out that someone was pregnant before they’d told anyone, and possibly even before they knew themselves. It was just based on data they had about how pregnant people shopped.
Sure, in theory a human could catch that sort of thing, but AI can sort through a lot of data quite quickly. If humans did it, they’d still likely need digital assistance in collecting said data.
For the record, it was Target. And, that’s the example always quoted, not just in the media generally, but also by vendors pitching their analytical services to mid-size ecom businesses. “Imagine what we can surface for you!”
(Yes, in this space, surface has evolved to become a new transitive verb.)
This is surprising, since it is a very dated example. Given improvements on technology over the decade, one would think insights to be even more profound. Companies send millions of ads to millions of people every day. Is it possible even this “proto-example” could be explained plain luck? In the ER, most women who could theoretically be pregnant get a test before many X-rays, and the number of positives is not negligible, which is why they do the test.
I’m wary of personalisation in general, for reasons abundantly made clear by the other posters. However, for streaming music I do believe personalisation may be useful. I noticed that music streaming services are moving towards personalised playlists, and on some occasions these did introduce me to new music and artists that I didn’t know before and liked. Granted, I have to wade through lots of recommendations that simply duplicate what I already know.