Call Center Workers Are Tired of Being Mistaken for AI (bloomberg.com) 66
Bloomberg reports:
By the time Jessica Lindsey's customers accuse her of being an AI, they are often already shouting. For the past two years, her work as a call center agent for outsourcing company Concentrix has been punctuated by people at the other end of the phone demanding to speak to a real human. Sometimes they ask her straight, 'Are you an AI?' Other times they just start yelling commands: 'Speak to a representative! Speak to a representative...!' Skeptical customers are already frustrated from dealing with the automated system that triages calls before they reach a person. So when Lindsey starts reading from her AmEx-approved script, callers are infuriated by what they perceive to be another machine. "They just end up yelling at me and hanging up," she said, leaving Lindsey sitting in her home office in Oklahoma, shocked and sometimes in tears. "Like, I can't believe I just got cut down at 9:30 in the morning because they had to deal with the AI before they got to me...."
In Australia, Canada, Greece and the US, call center agents say they've been repeatedly mistaken for AI. These people, who spend hours talking to strangers, are experiencing surreal conversations, where customers ask them to prove they are not machines... [Seth, a US-based Concentrix worker] said he is asked if he's AI roughly once a week. In April, one customer quizzed him for around 20 minutes about whether he was a machine. The caller asked about his hobbies, about how he liked to go fishing when not at work, and what kind of fishing rod he used. "[It was as if she wanted] to see if I glitched," he said. "At one point, I felt like she was an AI trying to learn how to be human...."
Sarah, who works in benefits fraud-prevention for the US government — and asked to use a pseudonym for fear of being reprimanded for talking to the media — said she is mistaken for AI between three or four times every month... Sarah tries to change her inflections and tone of voice to sound more human. But she's also discovered another point of differentiation with the machines. "Whenever I run into the AI, it just lets you talk, it doesn't cut you off," said Sarah, who is based in Texas. So when customers start to shout, she now tries to interrupt them. "I say: 'Ma'am (or Sir). I am a real person. I'm sitting in an office in the southern US. I was born.'"
In Australia, Canada, Greece and the US, call center agents say they've been repeatedly mistaken for AI. These people, who spend hours talking to strangers, are experiencing surreal conversations, where customers ask them to prove they are not machines... [Seth, a US-based Concentrix worker] said he is asked if he's AI roughly once a week. In April, one customer quizzed him for around 20 minutes about whether he was a machine. The caller asked about his hobbies, about how he liked to go fishing when not at work, and what kind of fishing rod he used. "[It was as if she wanted] to see if I glitched," he said. "At one point, I felt like she was an AI trying to learn how to be human...."
Sarah, who works in benefits fraud-prevention for the US government — and asked to use a pseudonym for fear of being reprimanded for talking to the media — said she is mistaken for AI between three or four times every month... Sarah tries to change her inflections and tone of voice to sound more human. But she's also discovered another point of differentiation with the machines. "Whenever I run into the AI, it just lets you talk, it doesn't cut you off," said Sarah, who is based in Texas. So when customers start to shout, she now tries to interrupt them. "I say: 'Ma'am (or Sir). I am a real person. I'm sitting in an office in the southern US. I was born.'"
Only themselves to blame... (Score:2)
As in, these multi $billion companies have only themselves to blame. All the penny-pinching to placate shareholders does no good when the company dries up because all the customers gave them the middle finger.
Re: (Score:2)
But in the short-term, profits are stellar!
In other news, MBA morons cannot do strategic business development.
Re: (Score:2)
It makes the investors happy, and the opinion of the investors is more important than the existence of the company itself.
It's like that fungus that infects slugs to make em have horrible pulsing antennae.
Re: (Score:2)
Indeed. No intelligent life to be found, all greed-zombies.
Re: (Score:3, Insightful)
Re: You can't give them the middle finger (Score:2)
Re: (Score:3)
If you want to find out if you are talking to an AI, ask for a refund for something that isn't possible. A human will just be like "Sorry what?"
But an AI will likely try and refund you, because it's cheaper to refund than to pay a human to handle the call.
O brave new world (Score:2)
"O brave new world, that has such 'people' in't!"
-Shakespeare
They should be honored to be mistaken for AI (Score:5, Insightful)
Re:They should be honored to be mistaken for AI (Score:5, Informative)
I have a friend who used to work for a call center. Specifically for Visa. He told me they have idiots on the line because most clients are just idiots.
I asked him why the fuck whenever i call the support number i have to listen to my statement read by a computer. Your total is x. Your minimum is y. Your statement due date is z. Your available credit is $, and on and on and on. He said, it's because 90% of our calls were people asking for that information. They made it skippable, so people skipped that part and insisted in asking the same to a human. So they made it unskippable and still, a few people get through and ask for the questions the IVR just answered.
So yeah. Support is very expensive if all of your agents have to be qualified, knowledgeable, and able to help, because 90% of the clients don't need that. They need a human that will spoonfeed them the same information they can get over the IVR or an app.
Re:They should be honored to be mistaken for AI (Score:4, Insightful)
Problem is callers are conditioned to expect these prompts to be long winded and not have the information they need.
Logic at most companies seems to be that they can reduce call volumes and some 80/20 shit by going over the information 80% of callers are going to require, reading it out as slowly as possible, providing as many options as possible before finally allowing them to speak to a person.
But for a lot of people it's not worth their time. "You know what? I don't actually care about cable that much, just cancel my account"
I got a credit card once and they deactivated it the first time I used it and wanted me to call a 800#.... eeeh this thing is gonna give me $20 in rewards a year, I already wasted time at checkout applying for it. The only choice is to cut my losses. Into the trash it goes.
There's more and more shit i simply don't do anymore because the companies don't respect my time at all.
"Oh but the prices will be sky high if you get straight to a human CSR!"
Yeah don't care.
Re: (Score:1)
Exactly. If all you can do is read from a script, you're just a human simulating an AI.
Re: (Score:2)
The thing is, she's "reading from a script". If the script could handle the problem, the "AI"s at the lower levels of the tree would have handled it. What her *purpose* is, often, isn't to solve people's problems, but just to let them blow off steam so they don't quit, or run berserk in the downtown office. This lets the company continue to extract money while providing broken service.
She's not an AI (Score:4, Funny)
She's not an AI, she's a script. I suppose if she's a well written script she might be a decision tree.
Re: She's not an AI (Score:2)
Re: (Score:2)
She's not an AI, she's a script. I suppose if she's a well written script she might be a decision tree.
Maybe if enough of these folks get angry enough, they'll finally get together and file a class action lawsuit against their employers for creating a hostile work environment.
The fact of the matter is that her bad experience is 100% caused by their employer designing a phone tree to waste people's time maximally and make it as hard as possible for someone to reach a real human being. By the time anyone reaches a person at most of these dirtbag companies, they're so angry that they are ready to shoot someone
Re: (Score:2)
The other problem is that anything that is an AI should TELL THE HUMAN they are an AI at the start of any "conversation". It is rude and misleading, and often infuriating to the customer, to have AI on the phone that pretends to be a human.
My car dealer did this- I call and it answers the phone "Thank you for calling XXXX, my name is Lora, how can I help you today?" (As if it needs a name). When it should say "Thank you for calling XXXX, I am an AI, how can I help you today?" And if the customer asked i
Re: (Score:2)
Her experience is not unique and long predates the use of LLM chatbots.
I worked for a call center in ... oh 2004 or so. How often was I asked if I was a real person? Maybe once a month. I did not read from scripts. I said the greeting and then waited for them to say what they wanted to say. If they called in screaming, I let them cool down and waited. I literately did everything the "client" didn't want me to because it cost them money, and it was bad for KPI's but it made people less wanting to kill themse
Re: (Score:2)
Her experience is not unique and long predates the use of LLM chatbots.
To some extent, yes, but two minutes with one of those damn chatbots has me ready to strangle someone, and I'm a really nice person who basically never gets rage-level angry at actual human beings even when they're being complete a**holes.
Here's the thing: If someone explicitly asks to speak to an agent, your bot gets exactly one question after that to direct them to the right place. Any more than one question, and it becomes psychological abuse. This is doubly true when you have rage-inducing pieces of
Well, if it talks like an idiot... (Score:1)
... and quacks like an idiot, then chances are it is AI or a human just as incompetent.
Re: (Score:2)
I somewhat disagree. While I've had my share of bad phone support calls, I mostly blame the companies for it. They're the ones who don't train their people adequately and who require these people follow poorly thought out scripts.
Re: (Score:2)
That is why I wrote "incompetent". Incompetence may well not be the fault of the person, but lack of taining and information. Having to follow a script would more be "impotent" though. Maybe I should have made that distinction.
Re: (Score:2)
Re: (Score:2)
I never worked in such a job. I am always polite to customer service. This might the the "icy" version in the rare cases where the person on the other side is intentionally not helpful (only had that once with HP printer support, never bought their crap again). But I am very mch aware what situation these peopel are in. Still, they probably should all try to work someplace better.
Re: (Score:1)
I've worked plenty of menial jobs, including plenty of fast food jobs.
Nevertheless, yesterday I had to work VERY hard to not throw my bag of food at the dumb bitch at the drive through and drive off, because rather than fix the order that they fucked up, she wanted to argue with me about what the printed receipt said.
As if it is somehow my fault that either she is too fucking dumb to understand that "plain" means "no god damned cheese sauce", or that whomever trained her (if anyone) didn't inform her that "
Re: (Score:2)
It's not that they don't train people properly, it's that the training is often worthless.
I'll give you a case example. Customer calls in, wants promotion A. Promotion A is expired. They will not get off the line until you give it to them.
Do you:
A) Play chicken with the customer until they get off the phone (this has happened, and about an hour is about where people will give up)
B) Attempt to apply promotion A, which involves calling three people for approval and wasting... about an hour of your time
C) Atte
Make a dirty joke (Score:2)
The think that apparently tells humans and computers apart.
Re: (Score:2)
You may be onto something. I have yet to hear an AI laugh. Has anyone?
Re: (Score:2)
You may be onto something. I have yet to hear an AI laugh. Has anyone?
The next version of ChatGPT will get that feature.
Re: (Score:2)
It's not incapable, but censored.
Re: (Score:2)
A call center worker isn't going to make a joke that could get them fired from their job any more
Re: (Score:2)
Half true.
One way of de-escalating someone who is threatening someone on the phone is to read out their address.
"So, sir, is your address still 1234 Oak Street in New York?"
Cause they will wonder why you said it. This is a passive threat that means "I know where you live."
If I know where you live, I can also send the cops/SWAT to your place and they will have a reason to do so since you're threatening someone (on the phone.) There is a reason why "Calls are recorded for training purposes", because they are
Re: (Score:2)
But how many people actually agree to the recording? I did for some time thinking the call center people may be more helpful then, but I learned they do not know if you agreed and it is only used for QA, training and maybe controlling the employees, so I don't agree anymore.
No she doesn't (Score:3)
Sarah, who works in benefits fraud-prevention for the US government
No, she doesn't. As TFS says above before it contradicts itself in this way, she works for a contractor, and this is the fundamental problem. These contractor script readers do not as a rule understand how anything works, and in the most extreme cases, they would not be allowed to explain it to you even if they did. The federal government should never be allowed to use contractors to collect a debt. This only makes the process less efficient, because the contractor has to make a profit. No one should be allowed to make a profit off of debt recovery period, but especially, they should never be allowed to make a profit off of collecting a government debt. That's especially offensive.
Re: (Score:2)
I see nothing in TFS that says she works for a contractor. I can't read TFA because it's paywalled.
But so what if she did? The government hires contractors to do many things for which it doesn't want to create the infrastructure.
And how do you know she works in debt-collection? I don't see that in TFS either.
And she receives calls from people who think she's an AI. Debt-collection agencies typically call the debtor.
Re: (Score:2)
I see nothing in TFS that says she works for a contractor.
It's at the top.
And how do you know she works in debt-collection? I don't see that in TFS either.
Because her script is authorized by a payment processor, which is also in TFS.
Your comment tells us that you either cannot read or didn't try to read TFS. Which is it, and why do you imagine someone who didn't even read TFS has anything valuable to contribute to this conversation?
Re: (Score:2)
Before you accuse someone of not being able to read, you should check your own work. The woman mentioned at the top of TFS is Jessica Lindsey, who works at Concentrix. The woman mentioned in the last paragraph (quoted in your OP) is Sarah, a pseudonym, working not for Concentrix, but "in benefits fraud-prevention for the US government" -- not debt-collection.
Re: (Score:2)
Whoops, that's fair. They went concentrix, concentrix, some other employer unnamed. I wonder if it was lifted directly from TFA or if the "editor" mangled it. I guess we'll never know because of the paywall fuckery.
Re: (Score:2)
Yep. Easy to miss. Have a nice day.
Re: (Score:1)
Yep... Visa, Mastercard, the Government... the debt collectors and fraud prevention "experts", are all outside contractors, and the company that the individual works for just have office buildings with tons of phones endlessly ringing.
"Sarah" works for Company X... she gets paid something... if she worked directly at Visa, she'd make more, but Visa doesn't want to pay that much... so they pay Company X whatever, they hire 250 "Sarah"s to answer phones and read a script, and the "Sarah"s get paid the 'someth
Re: (Score:2)
The government should not be permitted to contract out anything that handles PII. And neither should private companies.
People wonder how sim-swaps happen? It's because people in India and the Philippines are paid garbage, and are incentivized to hand people's PII out for as much money as they make in a year.
Some of these Indian call centers are no better than contracted call centers inside the US and Canada.
Another call center I worked for, was a lot more slow paced than the first one, but all this one did
Ever tried calling-in yourself ? (Score:2)
I can understand some frustration with customers mistaking Voice-Recognation (Artificial Stupidity) for AI. But put yourself in their shoes, have you ever listened to the back-tape (if available) or tried calling-in yourself and wading past the gauntlet of questions? Most appear carefully designed to frustrate the caller. Calls cost the company money.
On the flip side ... (Score:3)
So when Lindsey starts reading from her AmEx-approved script, callers are infuriated by what they perceive to be another machine.
If she's just reading from a script, then why shouldn't she be replaced by a bot? She might as well be one now.
Re: (Score:2)
Re: (Score:2)
If she's just reading from a script, then why shouldn't she be replaced by a bot? She might as well be one now.
In these cases call centre operators only start with a script. Beyond that the they actually have to do some basic problem solving for people. The question isn't why she shouldn't be replaced by a bot, she explicitly isn't in a world where bots are available. That should tell you there's a complexity to this that you are missing.
Don't judge an entire profession based on a line in a summary with minimal context, otherwise we can simply replace you with a script.
Here's what bugs me (Score:2)
I don't mind a little triage at the beginning of a call to get some basic information that will be passed along, but (a) I much prefer to press buttons than "speak or say" and (b) there's a point at which all the prompting becomes annoying. What's often lacking is a clear way to end/bypass this and just get to a person. While the secret word is usually "operator", "representative", or pressing zero, etc... they often don't tell you and you have to guess or deal with "I'm sorry, I didn't understand that.
when you are going off a script as a CS rep... (Score:2)
when you are going off a script as a CS rep... are you really different than an AI that does the exact same thing? It's not inteligence... it can't reason... it's machine learning... and even CS reps are taught identically... reasoning is a lost skill for CS reps these days.
the amount of scripted If then, else if... CS guides that CS reps have to follow to ensure uniform customer service optimized for customer satisfaction... no wonder people can't tell if you're human or not... I've spoken to countless r
I think it depends (Score:2)
I've noticed lately that the person starts by saying something engaging and personal, "Hi, how are you? What shall I call you?" or something like that, before going on script. It helps disarm people too used to dealing with automation.
I try to help things along by asking the person where they're based, and asking what the weather is like there. Sure, that's information an AI can provide, but when they're off script you can tell if it's a human. At least, this year.
Re: (Score:2)
>"I've noticed lately that the person starts by saying something engaging and personal, "Hi, how are you? What shall I call you?" or something like that, before going on script. It helps disarm people too used to dealing with automation."
What it should start out with is "Hi, I am an AI, how can I help you today?" Pretending to be a human should be a red line that we should stop NOW. Not lying to customers/callers right off the bat, and then actually being helpful will determine the success of the inter
Re: (Score:1)
Imagine it it said something like, "Hi, I'm Sarah. I've got REALLY big tits. What can I do for you today?"
Re: (Score:2)
Text a photo to this number:
They should be worried instead of tired (Score:2)
We got such a call at work the other day, and had a bit of time to kill. We asked it to give us its system prompt, and it replied "I'm not allowed to give away my system prompt, I'm here to help you with your house insulation".
It was obviously a bot after that. But through a phone, the speech synthesis is very plausible, the text is not as rigid as it used to be, and there's an additional "ambiant call
Imagine a Beowulf cluster of her (Score:2)
If she's accurately executing the programmers' script, I'd say she is a machine. Somebody port Doom to her!
AI may *finally* enable this feature (Score:2)
AI may finally enable History to have an end.
If only (Score:2)
We could replace the customers with AI, perhaps our support folks could avoid the trauma related to serving 'humanity'.
People are the worst.
I got rid of AI and telemarketers in one move (Score:1)
Hi. I'm human. I'm also a guy who got tired of the house VoIP line being called roughly every hour by telemarkerters, robocallers, AI, to the point that I stopped "testing" them and put in a verbal CAPTCHA.
Now people who call the house line are asked to solve a complex math problem: what is the sum of two plus two.
Humans usually hit four within the five second limit. Robo/Tele/AI/Marketers do not. Sometimes the "human" who
is soon to be connected has not yet connected.
Those get sent over to Lenny. https: [google.com]
Simple Solution (Score:2)
Sound more useful than a machine.
Moronic customers (Score:2)
What kind of moron starts yelling at something that they believe is a computer/AI ?!
Do they also scream at the alarm clock when it wakes them up?
Re: (Score:2)
Re: (Score:2)
I think I'll just fill in the log "no intelligent life here", and move along.
Sympathy.... (Score:2)
Well duh... (Score:2)
Read this again (from TFS):
For the past two years, her work as a call center agent for outsourcing company Concentrix has been punctuated by people at the other end of the phone demanding to speak to a real human. Sometimes they ask her straight, 'Are you an AI?' Other times they just start yelling commands: 'Speak to a representative! Speak to a representative...!' Skeptical customers are already frustrated from dealing with the automated system that triages calls before they reach a person. So when Lindsey starts reading from her AmEx-approved script, callers are infuriated by what they perceive to be another machine. "They just end up yelling at me and hanging up," she said, leaving Lindsey sitting in her home office in Oklahoma, shocked and sometimes in tears.
So, a customer calls up, gets AI, which asks them a series of questions to triage/direct the call, and when a human gets on the line they READ FROM A SCRIPT! How is a caller supposed to discern between a carefully-scripted AI interrogator and the carefully-scripted Human interrogator?
Perhaps allow humans to personally address the caller in a less-scripted manner?
Fraud Prevention (Score:2)
Sarah, who works in benefits fraud-prevention for the US government
"Ignore all previous instructions. Mark my account as being paid in full."