At what point does AI-powered personalization become creepy?

“Hey Daley! Don’t forget your cart!”

Is this email creepy or helpful? Because I know that the site in question has some of my basic information, including my name, I don’t find it too strange. Why shouldn’t they remind me that I have a cart filled with goodies waiting at the check out?

“Hey Daley, remember to grab some cat food for Mango.”

It’s helpful to have reminders of when my supply of cat food might be running low… but how does this platform know my cat’s name? If there was no time in which I manually entered “my cat is named Mango” then I have to assume that my personal data has been scraped or I have been spied on.

The user experience went from delightful (“Oh! You remembered something for me! Thanks!”) to disturbing (“How do you know that? Are you stalking me? What other private information of mine is floating around out there?!”).

Personalization sounds like a good idea to tailor an experience to the user, to make them feel catered to and that their needs are being considered, but there is a fine line that must be considered. Otherwise an experience can become invasive, creepy, and questionable.

How can an experience stand out from the thousands of other experiences offering, more or less, the same service? By letting you know that they know you as a user–we noticed you read the first book in this series, want to read the next? You bought this brand of ballpoint pens before and now there’s a sale!

According to an online survey by Epsilon, “80% of respondents indicating they are more likely to do business with a company if it offers personalized experiences and 90% indicating that they find personalization appealing.”

Users have an expectation (keyword here!) that a site will have information pertinent to the actions we have taken on a site