It happens all the time. You're online, checking out products, maybe looking for a holiday, or researching home loans, when all of a sudden there's a pop-up.
"Live help is available! Chat now with a member of our friendly customer support team!" It's easy to dismiss these chat dialogues, but it's also easy to take them up on their offer to help. Only, if you do, there's something you should probably be aware of.
As editor Tom Scocca from culture blog Hmm Daily found out recently, these online chats aren't the conversations they seem to be.
Specifically, even though support dialogues usually look like a series of sequential messages in time – where you ask questions, then wait for a response – that's actually a facade, at least some of the time.
Why? Because on some live chat platforms – particularly commercial ones like these geared to customer service – support staff can actually see what you're writing as you type it, even before you hit the return or enter key, or click send.
In fact, as Gizmodo adroitly points out, in circumstances like this, "the send button is an illusion".
Scocca stumbled upon this discovery after chatting with a customer service representative on a website that offers various discounts.
After taking some time to edit and craft a specifically worded question, Scocca was stunned to receive a written response to his query – complete with a product suggestion and a link – just one second later from the customer service rep.
"When the answer popped up, I felt off balance. It was too uncannily quick," Scocca explains.
"I remembered that I had, for whatever reason, been tinkering with my message before I sent it. Apparently my thinking about how best to present the question was moot, because the agent had to have been reading both versions while I was typing them."
Researching how this could be possible, Scocca came across a chat platform called LiveAgent that spruiks this creepy technological feature as a selling point for its software.
"Before the customer clicks the 'Send message' button, you have a chance to see in real time what the customer is typing," the company's site explains.
"This gives you more time to prepare an answer or solution to the customer's problem. Customers will appreciate your quick and precise answers."
Hmm, well we might appreciate it if we weren't so busy being creeped out by it.
Not that LiveAgent should be singled out for this voyeuristic practice, though – it looks like a number of commercial chat platforms have had the same ability for some time.
And, look, it makes sense. When customers want help, most probably want it quickly – if not unreasonably quickly – and don't want to waste their own precious time waiting around for support staff to answer their questions.
But at the same time, how many people are actually aware that this creepy casual snooping goes on all the time on thousands of websites across the world?
It's not like online chat apps advertise that they're actively surveilling the customer like this, and it's a strictly one-way window: we can't see what support staff are writing (although we're often informed if they're actively typing).
In these privacy-conscious times – where people are advised to lock down all their privacy settings – it just seems like a very weird blindspot we should all be aware of.
In a sense, these companies can monitor our on-the-spot thinking process when we turn to them for help, all thanks to the keystroke-logging magic of JavaScript. Nothing you type into that little box is sacred, folks.
Of course, real-time awareness of your innermost thoughts might not seem too intrusive if all you're actually wondering about is which dishwasher to buy.
But if you really want to keep that soul-searching to yourself, maybe draft any questions in your head beforehand, and not on your keyboard.