Online customer service agents are watching what you type


#1

Originally published at: https://boingboing.net/2018/11/28/online-customer-service-agents.html


#2

Watch me correct my typos in real time? Sounds like a fun job. /s


#3

Unlike just about every other entity deploying creepy applications of technology; customer service is an industry I can believe is simply too disinterested to use their surveillance powers to listen in on me. So there’s that.


#4

So let’s all start composing in a local text box and then copying/pasting in to chat box when ready. I expect there are a few apps for that … :wink:


#5

Any text editor will suffice if you want to do that.


#6

K…i…s…s…m…y…l…i…l…y…w…h…i…t…e…a…


#7

Which was entirely my point. (I should have added a snark tag?)
Ironic use of “apps” too. :wink:


#8

My girl was telling me this when she got a new position that requires her to use the SalesForce app.

Pretty much they never tell customers they can see it, but they do preliminary searches when they think the customer is going to ask a question so they can be a little more prepared.

Kinda freaked me out because there have been times my cursor is lost on one of my screens and I forget that I’m not currently entered into the private window where I’m doing my GGILF searches. GGILF WITH ONE BOOB DROOPING TO HER KNEES (Don’t judge! It will happen to all of us!!! And if you want to send me that photo when it does, I’m not going to say no!!!).

Nope, I don’t want Sales Force finding my porn and sharing it with liveagents!


#9

when we’re on the phone on hold with customer service, even as we listen to hold muzak, what we do and say while we wait can be heard and recorded.

And that’s why I have the mute button enabled on my local phone device even if I’m on hold.


#10

I worked in a call center. every inbound and outbound call can be recorded. Mute or hold doesn’t interfere with voice over ip recording.


#11

Facebook sees and saves your unpublished, self-censored posts


#12

I take the opposite approach and purposely say things that I’m hoping someone in customer service can hear. It’s mostly feedback on how terrible their phone menus, repeated ads, and music choices seem to customers who are on hold.


#13

Good. It’s time for a new spin on this scene from The Martian:


#14

That would work if customer service interfaces weren’t expressly designed to discourage/deflect and dismay the caller.

So your tips might cause them to think revise the interface downwards if the goal is to increase the number of people who give up and don’t take up the time of paid staff.


#15

Good point. I guess my experience with service-oriented companies made me think about improvements. In many industries, it’s become a race to the bottom - where businesses don’t seem to care if consumers choose a competitor.


#16

I get that “being on hold” by the company means that, in fact, they can still hear my voice (even as I hear their on-hold jingling music). But you cannot be suggesting that the company can hear my voice if I put them on hold and/or mute them on my side. (For a physical phone, on my side, at least. If we’re talking about a computer application, I agree anything is possible.)

The shitty trade-off is that as you type, you’re being observed, without your permission. Hitting a chat’s “send” or “reply” button is a universal sign that what you’ve got to say is ready to be seen: You’ve gathered your thoughts and have explained your needs to the best of your ability. That a customer service agent can see what you’re trying to say, before you actually say it, is rude and intrusive.

How shitty a tradeoff is this? What privacy do I really need in this situation? Isn’t it incredibly helpful to me (and, of course, to the call center) to be able to get the jump on all the common questions I might ask? Instead of me waiting seconds or minutes for a response (and perhaps hanging up), I do in fact get a timely response.

The tradeoff is, yes, they’ll see my real-time typing. I think really the only real safeguard that should established is that Cmd-V should be disabled in the chatbox (preventing accidental pasting of personal information or a link that the user thought they were pasting elsewhere). Otherwise, what I’m typing in the chatbox is relevant to the company and the sooner they see it, the sooner they can act on it. They are literally there to help me.

Of course, There should be a disclaimer in the chatboxes’ Terms of Service, buried somewhere (so that paranoid types don’t go nuts), and that should settle it.

I think it’s a fairly non-shitty tradeoff, provided there is disclosure.


#17

I have often wondered how many terrible jokes bbs watches me write before I settle on one
I have often wondered how many terrible jokes bbs watches me type before I finally just hit cancel


#18

Could be fun if their chat bot accepted all sorts of the old VT100 commands that BBSes used to - there used to be fun macros to spin a propellor (a long string of ‘/’ and ‘’ separated by backspace characters. Send a couple of hundred of those… or a bunch of cursor commands to move things all over their chatbots window, etc.


#19

Yeah - there must definitely be a way to leverage this information.

OMG I used the L word I swore I’d never use the L word I used the …

>cough<


#20

Disclosure is key, but I’m not entirely happy with the trade-off. By the time I am at customer services something has gone wrong and it is usually a complex or arcane issue. (It continues to amaze me how much customer services is designed to give only ‘subject 101’ answers that any user would have figured out - or ought to have done - long before they wanted to subject themselves to the horror of contact with the company.)
Given how many of the CS reps are not in my country (UK) and do not have English as a first language, I take care to try to be very specific and unambiguous about what I want to say. Even the slightest ambiguity or other lack of detail can result in “the 101 script” coming back in response, or an inaccurate or inappropriate answer, and the whole transaction takes far longer as a result and I have to type many, many more characters (and typing can be painful for me). So I do edit/revise as I type and before I hit send. Sometimes I deliberately want them to read specifically what I sent and only what I sent - so prominent disclosure by the agent at the start of the start of the chat (i.e. not in a weaselly small-print box outside the chat box somewhere else on the web page) ought to be mandatory.

Plus I sometimes have to tone down my comments before sending, given how frustrated and angry the fuckwit company and their badly trained agents have made me by the time I resort to contact and within a couple of exchanges where the rep has clearly not bothered to read what I wrote and has spotted a keyword that triggered a ‘script 101’ copy and paste response!