Slashdot Mirror
Customer Service Agents Might Be Able To See What You're Typing In Real Time (gizmodo.com)
Gizmodo is warning that some customer service agents might be able to see what you're typing in real time. A reader sent them a transcript from a conversation they had with a mattress company after the agent responded to a message he hadn't sent yet. From the report: Something similar recently happened to HmmDaily's Tom Scocca. He got a detailed answer from an agent one second after he hit send. Googling led Scocca to a live chat service that offers a feature it calls "real-time typing view" to allow agents to have their "answers prepared before the customer submits his questions." Another live chat service, which lists McDonalds, Ikea, and Paypal as its customers, calls the same feature "message sneak peek," saying it will allow you to "see what the visitor is typing in before they send it over." Salesforce Live Agent also offers "sneak peak."
This particular magic trick happens thanks to JavaScript operating in your browser and detecting what's happening on a particular site in real time. It's also how companies capture information you've entered into web forms before you've hit submit. Companies could lessen the creepiness by telling people their typing is seen in real time or could eliminate the send button altogether. So if you don't want to be monitored or send secret messages to agents, put your phone on mute while on hold and copy/paste messages from another document to your customer service chatbox. And in general, be nice to customer service agents. It's not their fault.
This particular magic trick happens thanks to JavaScript operating in your browser and detecting what's happening on a particular site in real time. It's also how companies capture information you've entered into web forms before you've hit submit. Companies could lessen the creepiness by telling people their typing is seen in real time or could eliminate the send button altogether. So if you don't want to be monitored or send secret messages to agents, put your phone on mute while on hold and copy/paste messages from another document to your customer service chatbox. And in general, be nice to customer service agents. It's not their fault.
Magic trick? Wha?
The logical solution would be to insert a false delay of 3-5 seconds whenever the user types something to the CSR's replies so that it is never perceived to be answering too fast.
As for the creepiness factor—you're typing stuff into a text box on a website, nothing should be considered hidden from them.
...
hundreds of 'nerds' reveal they've never seen or thought about how xml->http request systems work and don't think they should have to.
hilarious but something we all need to do now is be careful of what we say. never know who or what is recording sadly.
If you're talking about livechat, then yes, they absolutely can see what you are typing in realtime.
At least, with three different systems that I've used.
What are you worried about, that you'll type "go fuck yourself" and then delete it? Just don't do that.
You are going to take away an angry persons filter without telling them!? This is supposed to HELP somehow? Maybe the dumbass is the person who decided this was a good idea.
This should be common sense to anyone using a search engine too, it sends your requests keystroke by keystroke to update the suggestions. It's just how most JavaScript is coded for that type of thing.
One way to get around this is to quickly type what you want first in Notepad and then quickly copy/paste it into the chat or wherever you don't want to risk typing any mistakes to.
What's the problem?
I always assumed that the CSRs could see every keystroke I type in the chatbox.
If I have a long question (often the first question is long for example), I usually type the text into emacs (and, if I'm being picky, spell check it as well) and then cut and paste into the chatbox just to avoid confusion (such as me leaving out a "not" and later correcting it and the CSR never noticing my correction).
Why is there an "insightful" mod and why isn't it "-1"? If I wanted insight, I wouldn't be reading
I just chatted with Amazon support yesterday. I'm gonna type, "Fuck you," and wait for a ten count and then back over it.
It little behooves the best of us to comment on the rest of us.
If this really bugs you, just type your response into some other text editor. Notepad, or even a comment submission window on slashdot. Then cut and paste it into the CSR window.
This is probably happening. Just as likely is the person has a million calls about your issue and knows what you are going to say. They likely have some canned answers. I'm on the other end and I have canned information; I tech support chat so often I have a txt file with everything in it - explanations of troubleshooting processes, error codes, contact information - I know what they are going to ask me. Like I tell my kids, I'm not smart, I'm just old...
I had this same experience. He answered a question before I hit send.
If you're playing Customer support roulette, you only have to type it once and keep cutting and pasting it as you're handed off to another CSR.
Maybe save in another window:
If this isn't fixed today, I'm going to drive down there and exercise my second amendment rights /s
Bar *wp.com* and that will take out much of that kind of tracking.
Old news. Very old news
"live chat" has caught up to ICQ in the 90's.
Back in the days of "smart" (and "dumb") terminals, live chat enabled some attacks.
For instance: sending a control sequence that reprogrammed a "soft key" and then "pressed" it, allowing the attacker to execute commands as if the victim had typed them. (This could include suppressing the visibility on his screen so he didn't know it had been done.)
I wonder if these systems have an analogous vulnerability?
Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way
Is this a big deal? I worked on a CSR chat system for a major clothing retailer back in the early 00's that had that feature. If I remember right, there was an ActiveX version for IE and a javascript version for anyone else and you configure it to see as the user typed or wait until they hit "Send". The idea was that if you watched the person type, the CSR could start to compose their answer so that there was a small delay as possible.
This feature could be used to game the system. Slowly type in some question and leave it for a few seconds for them to digest. Then make a minor change that changes the sense of your question and quickly hit send. If they don’t notice the change, there’s now a chat log in your favour. Profit.
I was chatting to my phone company about my modem. They lulled me into a false sense of security by asking for info I'd already given. But a bit later I thought better of a somewhat tetchy paragraph and replaced it with a calmer one. Then they referred to something in the paragraph I had not sent.
Then they offered to call me and I had a friendly conversation with a nice young lady who fixed my problem. I was left wondering: good customer service, or psychological warfare?
Bah! Old bulletin board systems (BBSes)'s had real-time chats too. Now, get off my lawn! :P
Ant(Dude) @ Quality Foraged Links (AQFL.net) & The Ant Farm (antfarm.ma.cx / antfarm.home.dhs.org).
I used to work in a technical support role. We used a chat service provider called comm100 which does show the agent what you type in real time. At first, I felt dirty like I was invading their privacy, but it does help efficiency considerably. It's invaluable when it comes to de-escalating clients that start to type out a giant rant, then slowly edit it to be more civil... eventually they just chicken out and delete their entire whiny post when they decide it's not worth it. It feels bad and dirty, but also consider that the support agent is usually multi-tasking between clients, so being able to know what you are typing as you are typing it is a real time saver. The worst is when the client has typed out part of a question which you know the answer to and have a full response typed out, then you have to wait for them to hit the 'send' button before you can continue answering the question or solving the problem. Sometimes you have to wait a LONG time.
In ~2006 I integrated an open source real time chat system into a website for an ISP. "Message Preview" is/was a standard and on-by-default feature in that solution.
Regrettably I don't recall the name of the software.
you type normally and it inserts random crap, then edits it to the correct keys on enter
Normally I am paranoid about privacy, but this seems like a good thing. Why wouldn't you want them to respond faster?
By including this sig, the copyright holders of this work or collection unreservedly place it in the public domain.
'nuff said.
The chat client ICQ was doing this back in 1998. Why is it surprising that this would still be possible 20 years later?
*slight crashing sound*
https://youtu.be/v_UyVmITiYQ?t=620
I've already had a couple of conversations with support agents when their replied to the messages I hadn't yet sent which made me feel extremely awkward but now I know that I should never ever paste anything in such chat windows without first verifying that my clipboard contents are the one I really intend to share.
The logical solution would be to make it symmetrical; if the rep can see what you're typing, let the user see what the rep's typing. Nobody would be surprised then. The problem is caused by the asymmetry, not because you're seeing unfinished messages.
At any rate, it is so easy to get around it, if necessary, that it hardly deserves any discussion.
Sometimes it is their fault, like when they're trying to do something unethical or illegal with the excuse being that it's part of their job. Fuck them. Fuck them sideways. They're part of the problem. Getting paid for your bad behavior doesn't make it less bad.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Evil technology!
Why, back in my day we called people, and of course the person on the other end of the telephone doesn't hear anything until I finish a sentence... oh wait.
In the "old days" we had a program called talk that it was apparent what you were typing appeared in realtime.
Also, this is exactly how REALLY old school text communication, TTY/TDD (for deaf/hard of hearing people) work. The text goes out as realtime. It actually works better than sending a whole composed sentence.
With only two people on the line at a time, it makes sense, and can lead to faster communication.
yes we see your posts as you type them. Sometimes I want to reach through the screen and hit enter. I stopped pre responding to unsent posts, because they can and will change. Don't embarrass yourself, we do watch what you type. If your wary about entering your contact information, don't type it in then delete it. If we're paying attention, we can capture it.
God I hate my job Please hire old programmers for something other than cold calls and chat rooms.
Someones probably reading this post before I send.
I will come to your HOUSE and BURN your DOG down!
There's obviously no reasonable expectation of "privacy" in this scenario -- but frankly, if you're using a computing device of any flavor, and if that device is attached to the internet, then you should probably just automatically assume that anything you do on that device could potentially be tracked. It obviously won't always be the case... but it might be the case, on occasion. This thought process kind'a falls under the "plan for the worst, hope for the best," kind of thing.
As an obvious direct example: I was attempting to negotiate with Verizon just recently, to try to circumvent an obvious "bait-and-switch" tactic that they had used during the Cyber-Monday promotional. At one point, I think I made the mistake of typing a comment into my "open mic" chat session, that essentially admitted that I was on the verge of just taking the deal, regardless of whether or not they agreed to give me what they had originally promised -- but of course, I backspaced over that before submitting. I realized a few minutes later what had probably happened, as my leverage completely evaporated in very short order. That mistake effectively cost me $200. (I left an absolutely scathing review of the CSR after the fact, describing what had happened... which, in retrospect, probably prompted some enthusiastic high-fives and kudos from the CSR's manager. So yeah... fuck you, Verizon.)
Lesson learned, though: just always assume that the mic is open.
If this was disclosed at the outset then no one would have cared at all. The problem is that we live in a world where "ask for forgiveness, not permission" is the default way of doing it even though individually we all know how wrong that feels.
I wonder what other javascript 'mickeys' like this we need to watch out for. Having something like this without disclosure of what is really happening borders on malware.
So all the send button really does is make the user text appear in the main scrolling chat window. A bit like those 'walk' buttons on traffic signals that beep when you press them, but you still get a walk signal regardless, and there is no change in wait time or duration of signal. It's just there to make you feel you are doing something.
...if a user seeking tech support is super frustrated, types in "FUCK YOU! FUCK YOUR COMPANY! ALL OF YOU GO FUCK FARM ANIMALS..", but then cools down, backspaces over all of that crap, and types in a rational query/response? Yes, the agent still sees the nastygram even though the user didn't think he sent it at all (and rightfuly so as nobody expects anything to go over the wire till he/she clicks "send")
What could possibly go wrong?
Facebook have had this 'feature' for posts you make, edit or delete forever.