A majority (64%) of 5,728 people surveyed by Gartner in December 2023 said they would prefer companies not to use AI in their customer service. Additionally, 53%...
If it worked for most shit and escalated to a human when it actually needed to, reliably, I'd be fine with it.
I don't believe there's a realistic chance that there's a lot of overlap between the people willing to invest to actually do it properly and the people paying for AI instead of people though.
The problem is the same as with the telephone answering trees.
If they’re used to help you get where you’re going, then they’re great. But that’s not the best financially motivated decision. Solving your problem costs the companies money. Pissing you off and convincing you that your problem shouldn’t be fixed saves money on support.
So making you go round in circles is the machine doing EXACTLY what they want it to do.
I get one of those meal kit delivery services. Every few weeks I'll go to their AI customer support and ask for cancellation and it'll give me discounts on upcoming orders. I keep the service at about 40% off at all times. Also when there's a problem with the order the chat bot just tosses me a discount. Cases like this are perfect for AI customer service.
Edit
Wow this blew up in a weird way. Just to be clear on a few points:
With the discount I pay $87 Canadian which is $76 untaxed or about $55usd. I also pay for this service using gift cards from Costco that are 20% off ($100 for $80) bringing that $55 weekly cost down to about $44. For 6 different dinners for me and my wife delivered to my front door every Monday. With crazy grocery prices where I live I cannot come close to beating that without giving up something. I won't eat the same thing every night (Sunday meal prep bros, don't at me), I don't want to expend the mental energy gathering recipes and ingredients but I do enjoy cooking a lot. It's something at the end of the day I can do with my hands free of screens. At regular price this was worth it to me, at 40% off it's actually saving me money. If they're still making money shipping this big box off food to me on a weekly basis, then good for them, we're both coming out on top.
Except they're selling you the kit at waaaay over cost in the first place, so they're still making money off of you. I promise you they are aware of the "glitch", and are not ignoring it out of the kindness of their hearts.
(not criticising you for using the service, if it works for you go for it and get those discounts, but don't let them manipulate you in to thinking you've got one over on them, they 100% account for this kind of thing and are still making money)
In my experience the AI assistant is just trained on the information available on the firm's website.
In 2024 I never just call a company expecting to be able to be assisted by a person. It's always quicker and easier to figure out how to interact with said company online. The only times you call are when it's not possible to resolve your query by interacting with them online.
That being the case, the entire purpose of the AI in this case is just to make it less convenient to call them. "Have you tried to resolve your issue online? Are you really sure about that? Maybe I could paraphrase this blog post from our website written by an intern 12 years ago."
90% of people calling support lines are due to questions that are in the top 10 ten on the FAQ. They're just the type of people who don't like reading and just want a social answer. The same kind of people who get told "just do a search, this is asked weekly" on Reddit.
If there was a way to direct the "I just need a FAQ that I don't need to read myself" people to an LLM and the "something is actually broken I need real help" to people, that would be ideal.
Repeat these words over and over. Most automated phone systems are programmed to bail out when its clear the customer is just flat out unwilling to engage with their bullshit.
I usually use the "cuss at the bot" method. Gets out my frustration ahead of time so i can be sweet with the human. Tho one time the computer hung up on my ass haha
I've called companies that disconnect the call or "in order to connect you to the right agent, please tell us what you're calling about," them inevitably get it wing enough times to make you sit through a menu of about ten choices that are not correct and disconnect after three rounds of this nonsense.
I extremely hate this idea. I I already hate the automated systems that are definitely designed to make you give up just trying to talk to an actual human being. Hopefully, we can get more lawsuits around the world like the Air canada one where they are liable for any bs the ai decides to make up, along with actual laws saying the same. Hopefully, it would discourage them.
I had the displeasure of being called by one from a vendor. It pissed me off that they couldn't be bothered to pick up the phone and call using a human, with how much we paid them. I canceled that contract and went with a different vendor, and let the sales team know exactly why. LLMs have their place, but my time is not the waste bin.
Hahahahaha!! I was sitting there, on the Pick Username screen for a good 5 minutes, singing that song in my head, trying to think of a good username. After a while, I thought to myself, "that's a good enough username, in done thinking about this", and sang it out loud as I typed it in....... 3
The point of modern "customer service" is to NOT provide customer service. If you can drag out the conversation to the point where the caller rage-quits in frustration, then the company can avoid spending any money on fixing any problems they've caused.
Previous way for companies to cut down on customer support costs was to make a better quality product (making support interactions rarer). That is not so much the philosophy anymore.
It's also similar to scammers. When you are not quite certain if you've been scammed, you'll first ask. There's a percentage of cases where you won't bother for the sum, because you've used the energy on pinging them.
While in case of companies you could have used that energy to, say, post "X is crap" somewhere in the Web.
Already out there in certain ways. There's a restaraunt near me that uses an automated system to collect orders in the drive-thru, and puts them into the system incorrectly.
At least that's what seems to be its purpose, because it does that really well. That, and piss people off.
Storytime! Earlier this year, I had an Amazon package stolen. We had reason to be suspicious, so we immediately contacted the landlord and within six hours we had video footage of a woman biking up to the building, taking our packages, and hurriedly leaving.
So of course, I go to Amazon and try to report my package as stolen.... which traps me for a whole hour in a loop with Amazon's "chat support" AI, repeatedly insisting that I wait 48 hours "in case my package shows up". I cannot explain to this thing clearly enough that, no, it's not showing up, I literally have video evidence of it being stolen that I'm willing to send you. It literally cuts off the conversation once it gives its final "solution" and I have to restart the convo over and over.
Takes me hours to wrench a damn phone number out of the thing, and a human being actually understands me and sends me a refund within 5 minutes.
My guess is you're one of the 10% or so who didn't give up in frustration. My % assumption might be off, but assuming any percentage of people gave up and walked away without costing Amazon a dime the system was working perfectly.
Dude could save yourself time by just going to contact page and ask for a call. I never use these companies chat features.
Also I found if I Google customer service numbers regurdless of company than I can get a number to call 85% of the time.
Of course after that you either got to fight robot to get a human on the phone that 9 times out of 10 will be a person out of India who also acts like a goddamm robot that doesn't understand English.
But my biggest pet peeve is a lot of times I have ro get a supervisor to solve a problem that would take the customer service agent ten seconds to solve.
Historically, these chat interfaces were tied out to a call center somewhere on the opposite side of the planet. Now they're entirely prompt-engineered. So you used to be able to work a claim through chat without sitting on a phone call for hours at a time. But now they obscure their customer support phone number behind six layers of tabs and links, while shoving the "WOULD YOU LIKE TO CHAT WITH A REPRESENTATIVE" button in your face the whole way, fully knowing it doesn't actually connect to anything that will help.
But my biggest pet peeve is a lot of times I have ro get a supervisor to solve a problem that would take the customer service agent ten seconds to solve.
A lot of the agents are just working off of written prompts anyway. But they do get experience with these problems over time (or recognize a slew of the same problem coming in at once) and can cut through the shit to give you a real, human response. Sometimes that response is simply "We can't help, because of widespread technical / systems issues", but that's better than being bounced through an automated service that feeds out generic non-answers and useless how-to guides.
So of course, I go to Amazon and try to report my package as stolen… which traps me for a whole hour in a loop with Amazon’s “chat support” AI, repeatedly insisting that I wait 48 hours “in case my package shows up”.
I tried to change the dates of a car rental through Priceline, a day after I entered the order. I got a message saying "You cannot change this order until 72 hours before your arrival" which I thought was weird. But I bookmarked the date and called as soon as I was inside the window. "Oops! Sorry, you can't cancel or change the reservation because too much time has passed!" was the automated response.
Absolute fucking scam. So I submitted a complaint through my credit card company to reject the charges. In this particular case, automation worked in my favor, because AMEX's dispute process is as opaque and arcane for the vendors as Priceline's support desk was for its own clients.
But its increasingly computerized horseshit. Nothing actually fucking works, except the vacuum they hook up to your bank account every time they find an excuse to extract payment.
The market does determine, unfortunately the market is relatively unfazed by subpar customer service. It has to be really bad or a huge legal catastrophe before it moves the needle. Which is why phone trees and long wait times are ubiquitous despite being universally hated. Marketing and sales and having a 90+ % rate of people that don't ever feel the need to call customer service basically eliminates that bad service as a concern.
Even when asus had a famously bad customer service scandal this year, their sales continued to rise unabated.
Companies don’t want to provide actual service for problems. That costs money. They want you to give up.
Customers hate anything that actually gets between them and someone that can actually help. Not shitty, complicated automated phone menus. Not some underpaid stooge who refuses to da anything except read from a mandatory customer service script. And not AI, which will combine both of the worst aspects of automation and scripted service along with a cheerful idiot that will spare no effort to direct you away from the nearest actual assistance.
I do like it in the sense that people HATE working in customer service. Because people have zero respect and customers make your job day miserable all the time.
Is one of the places where people deserve getting a hallucinating robot as a vengeance for how bad they treated people that worked there.
I talk to about 10 customers each day for no more than 30 minutes and 99% of the time it’s to fix something they messed up on. 30% of those people are jerks.
Thankfully most of my job is NOT dealing with customers and I truly feel bad for people who have to deal with them in high volume each day.
If I call them, I can fix it immediately, if they call in, enjoy the robot. So don’t be mean and the call won’t be disconnected (which I have permission to do fortunately)
Tangential, but I absolutely loved working in technical support. The satisfaction of actually helping someone with a problem affecting their real life totally outweighed the abuse from individuals who were letting the work part of their life drag the whole rest of it down (which was just kind of sad to watch). I've gotten paid much more for other roles since then, but it's one of the few roles in which I was thanked for what I did by the person I was working for, and that makes a huge difference.
There's this boomer obsession with making it listen to human speech...
Nobody under 40 wants to use human speech to talk to an AI. We don't want to us human speech to talk to humans most of the time, especially if we don't know them.
But they always want to jam an AI into areas where human speech is the main communication method.
The absolute last place AI should have been deployed is answering a phone call. Because that is the last resort for most people, but the boomers calling the shots still think that's people's go to move before trying anything else
While some of this is cultural, it’s also about accessibility. Old people want to use their voice because their sight is often less reliable and they aren’t as good at pushing the right buttons. My father for example is functionally blind and voice is all he has. So before we get mad at boomers calling all the shots, let’s consider that they’re not just old fashioned. They’re old. and so will you be one day.
Automated phone systems have been a thing for decades. They are notoriously shitty and adding a layer of “friendly AI” on top of that shitty system doesn’t bode well.
and that's even before it has to contend with you having an accent, or the mic quality being anything less than crystal clear, with a perfect connection.
Realistically we only dislike it because it’s a half baked solution. I know that if those LLMs actually did anything useful we wouldn’t mind them. But all these LLMs do is spam the documentation, which is already on the vendor website anyway.
There’s a NYT article somewhere, and I’ve been desperately trying to find it, about a woman who worked as some kind of real estate(?) call center AI augmenter. Essentially people would call in about listings or something, and she had to step in when the AI went off the tracks or didn’t know how to answer questions, matching its tone/inflection while refusing to acknowledge that there was a human stepping in. She ended up being super burnt out from the job. So the whole system was just super redundant, awful for the people working there, and as we’ve come to expect from AI, just a half-baked turd sold to some MBAs for a mint.
A lot of "AI systems" are literally just a dude in India. Outsourced human labour is cheaper than building and running a proper AI driven system, but companies want to say they're "incorporating AI" because it makes shareholders dicks hard.
Are you normally amazed by people hating on environmental disasters which are being marketed as the great solution to the world's problems but are only actually useful in a few industries and not to the general public overall?
I'm already pissed with bots, had to call my ISP yesterday because my internet was spotty, I couldn't talk to a single human, the bot was walking me through the tired modem restart, and then it ended the call and asked for me rate it even though it didn't solve anything!
I worked for an ISP. These problems are rarely ever ISP problems. It goes like this. ISP offers 50Mbps–1.2Gbps. If you are a cheap bastard and opt for the lowest tier plan you get a cheap hardware and if you don't ask for an upgrade you'll run that box until it doesn't work. So you have people rocking hardware that was manufactured in 2009 and installed in 2014 wondering why their cheap ass WIFI4 box installed in their basement doesn't work so well in half their house in 2024.
What's more they have a download speed that would have been good in 2009 only instead of 2 computers they now have 20 connected devices and stream in 4K.
What's worse is the rental on that shit WIFI4 box is about $20 a month or $2400 over 10 years so your paying for a BMW and getting a Pinto.
Smart people buy their own access points preferably wifi 7. Get one per story of your house and connect them with a physical Ethernet cable. Arrange them so that they overlap but not that much so that you don't have dead zones. If you work from home get a proper desk and run a physical Ethernet cable to your device. Also if you have devices that are literally 2.5 feet from each other and they support physical network cables just plug them in. Don't be that guy spending an hour trying to figure out why his router and his printer/tv aren't friends when they are almost touching each other.
So the ISP isn't to blame when the cheap ISP-provided hardware fails, and the solution isn't for the ISP to replace insufficient ISP-owned hardware but for you to buy your own instead?
The "wire everything" approach is a little excessive for most home networks too, outside of exceptional circumstances modern WiFi on modern hardware is more than enough for home users. It's only worth the time and money to wire everything if you've identified specific issues with signal loss or noise, don't just do it by default.
The hardware the ISP provides is always an ISP problem. Provide hardware that actually works.
Also, unless you're fiber, you don't provide the bandwidth you actually sell people, which is also an ISP problem. Every single customer who can't get their advertised speed at peak load should be a mandatory criminal case of fraud.
That's cool and all but it was a regional problem on their end, I learned after trying the whatsapp bot, which worked way better than the phone piece of shit.
I actually had a pretty godawful hardware provided by the ISP years before, that I just killed in salt water and said it wasn't working, then I got a new one from them that actually had a good wi-fi range :).
The other day I was able to take care of what I thought was a reasonably complicated customer service issue through an automated assistant.
I take a daily prescription medication and it’s on automatic refill. However now and then I forget to take my pill and then I have an extra. After years of this I found myself with 20-30 pills left when my next bottle was ready.
So I tried to call the pharmacy and say hey that bottle of pills you have waiting for me? I still need it, but not for about 3 weeks. Can you push my entire schedule back that much but otherwise keep the pace the same?
Turned out I was able to do this just by listening to menus, selecting from multiple choice, and entering numbers for dates.
I was so satisfied! I don’t want to talk to a human if I can possibly help it. I’d much rather deal with an automated system as long as it can do what’s needed. The problem is that most of them can’t. But then again most customer service humans are useless too, so…
It would be so much better to just have a website with all of those options. Why does it need to be a phone system at all? I hate calling in, I just want to enter an identifier of some sort and click a couple buttons...
Yeah that’s an excellent point. Older generations still prefer phone calls but I imagine increasingly the only people who want to call will be the people who can’t fix their issue via an automated system.
Because they are not trying to design an efficient system. They are trying to design a system that is as cheep as possible, puts off as many customers as possible from interacting with it while not being so bad as to fall foul of regulations. A well designed website that efficiently took you to the right place to make your complaint and get money from them/make them do something would fail requirement 2.
Around my way, we have a pizza chain where they've began utilizing AI to take orders over the phone. The only screw up the AI made was that at first, before the process of taking our order down, it wanted to confirm that we live within the delivery distance, so we provided our home address and it verified that we were within range of delivery, after taking the order and repeating it back to us, including that the order will be delivered to our home address (providing the details of the home address) within a certain time range, the moment it asked us if this information is correct, we said yes and then a long pause, and it responded that it could not verify our home address.
Wat.
And because we decided to speak to a human, it apparently dumped the entire order and the person who answered our call did not have access to all the details we provided the AI.
Pretty much wasted a little over 5 minutes with the AI.
Me...it's literally cheaper to call then use the internet to order. Try comparing the in-house menu, to the bullshit apps, to the website menu to calling and asking for a deal.
Calling is always cheaper especially if you pickup.
I also refuse to use any automated system. 0#0#0#0#0#0# or I keep saying human, representative, human human until the shiity programmed not gives up. Worst case, I actually go to the business in person.
The internet and companies is broken beyond belief.
Yup. I was literally born in India, lived there until I was 7, and have an Indian mother who very much still sounds Indian, and even I struggle to understand what outsourced Indian/Pakistani call centre staff say sometimes, especially when there's background noise.
And there's almost always either background noise or a bad connection. Sometimes I go sit in my car and listen over my car speakers, which are decent speakers, and it doesn't even help.
Yeah it turns out that using a statistical model to handle customer service leads to a degraded customer experience, because statistical models aren't humans and lack many human attributes.
The telephone services i had to use in my Lifetime were so insanely bad, thats one of the few things llm could do better than an underpaid person who has no will to live anymore and got yelled at for hours. This shit needs to end.
No. What actually needs to happen is for companies to give GOOD FUCKING CUSTOMER SERVICE!!! Try getting ahold of a real person at Amazon for example. I ended up cold calling the delivery company that handles local deliveries to get a phone number to talk to an actual person.
The next time i called they had turned that number into a robo call center...
I never had issues with Amazon support. To be fair, the last time I needed support was over 4 years ago, but still. I got to chat with some real person relatively quickly who managed to address the issue swiftly.
Unless they hate it enough to ditch a business or service in great enough numbers that it costs the business more money than they save by outsourcing to a computer, people had better get used to it.
The problem is that consumers likely don't have that choice. The "free market" is really bad in incentivising good long term behavior, they favor short term gains for their stockholders. Thus they likely all switch to practices that seemingly lower cost or raise short term profits. If they can fire employees and replace them with AI, they will do so.
If they would think long term, they would prefer to hire humans instead of AI, because that way they would give their future customers money to buy their stuff. AI will not be their customer. They would pay them enough money to be a happy and good consumer.
Customer choice doesn't matter here, they either just have to buy whatever is cheapest, or die, because their employers (if they even have one) don't pay they enough for them to have choice, because short term profits.
Yeah - that's all part of the "unless enough people leave" point.
It really depends on the market though - if it's not an essential good, it doesn't need to be replaced (online games). If there's adequate competition, there's largely undifferentiated alternatives (utilities around me)... and if not, you probably don't have a choice (your local government services, monopolies, and shallow markets for essential goods).
I don't think customer support can be resolved by free market forces. If someone has purchased the product, has a problem, and is trying to contact support to resolve the problem, they're a bit too far gone on the model of free consumer choice, and that instance won't affect the free market.
I feel like we need legislation that, when a customer has a problem, they must be able to contact the company for a refund or resolution, AND, communication with an "AI" does not count as that communication.
While I’d prefer to just speak to a human, I’d much prefer AI over the status quo of dead dumb automated systems that just keep looping through the same preset options until you get enraged and give up or mash zero
I love the idea of useful and improved AI-automations, based on the fact every site currently has a “customer service robot” that have never once helped with a resolution at all.
IMO it can eliminate a huge amount of support queries and leave the important stuff to the actual agent if done right. …with the caveat of yeah fuck AI if it’s fully instead of agents.
AI can only give you the options it was programmed to give. A Human is able to actually think and find a solution or direct you to a solution.
Your options are less with AI for customer service.
AI works best for applications that it is tailored for. But expecting it to "think" like Humans do is so far off. AI is being fed so much biased information and that is not "thinking" or learning.
I went through a phase of making the ai robot agree with me that it was the "email flange" that was causing my issue before transferring me to an operator.
Exactly! If I am calling customer support, it's because I have exhausted all other options of finding a solution to my issue, and I have a feeling I'm searching more extensively than the options that this AI is being fed. If I've reached the point of calling, I need someone that can think of a creative solution.
They may hallucinate but they're less likely to lie for now. I can't count the number of times I saw or heard some bullshit being told to customers to appease rather than help. And it irks me to no end every time I get a rep like that.
I don't blame the workers. I blame the corporate bullshit that actively encourages it by dangling bonuses and taking away if a customer doesn't feel that their issue was resolved. Call centers suck ass for both customers and employees.
How harrowing it is to hear someone a few cubicles away scream at one poor review by the end of the month that lost them a bonus for some bullshit that was out of their control after enduring so much abuse.
It is now at the point where we need to ask how they plan on handling complaints and problems. And if the answer is not correct, go somewhere else. Up till now this was never something we needed to worry about
I wonder if it would actually materialise, consisting the recent case where an airline company's AI chatbot promised a refund that didn't exist, but were expected to uphold that promise.
That risk of the bot offering something to the customer when the company would rather they not, might be too much.
It seems more likely that companies will either have someone monitoring it, and ready to cut the bot off if it goes against policy, or they'll just use a generated voice for a text interface that the client writes into, so they don't have that risk, and can pack more customers per agent at a time in.
The only good thing about is that, with most companies, if you need a refund for something under $10, it generally just gives you the refund and sets you off on your merry way.
Fair point but no one wants to deal with my incessant whining, and you couldn't pay me enough to deal with it. Sometimes a scratching post is what the cat needs.
I would like it if it actually did something for me, (like automatically doing x, y, or z to my account on the backend based on my request) but instead it just feels like every one of the "AI-Powered" support bots is designed to try and make it as hard as possible for me to actually get anything done.
An exceptionally well trained AI customer service has the potential to be amazing.
I only call or try to chat/email with customer service if something has gone way wrong - like outside the typical customer service capability of assistance.
If an AI can realize that my problem is human worthy and escalate it faster, that would save me time in the chat queue talking with someone who barely knows my native language.
Alas, AIs will be poorly trained, so the bad-english CS reps will still be right behind the AI interface waiting for me.
Anyone who ever tried to solve any problem and gets stupid responses on ai chat instead of making it easy to reach a real person that can solve it in seconds knows the pain.
Honestly, I've used some pretty decent AI chatbots. They can help you with basic questions and contact you with a human for things that require it or if you ask for it. Chatbots that don't let you talk to a human on the other hand, those are awful.
Unpopular opinion and rant:
Us, the consumers, brought this on ourselves.
Not intentionally but it was a slippery slope.
No one I know did ever ask the sales representative "does your customer support answer within 5 minutes and will I always reach a representative with att least 10 years of experience, that has the authority to make real decisions?". No, but we were all very interested about the pricing of the service/product.
Then these "Please press 1 for...." happened.... and no one of us really cared about the change because the service providers offered a much lower price than the ones with customer support representatives with 20 years of experience.
Since all of us went for the cheapest provider, the other ones had to cut cost to be able to offer their service on a competitive price level.
So then there were no one offering competent support with representatives that knew their shit.
And it slowly continued to go downhill...
So here we are with shitty services, which we pay for, where we all are treated as cattle.
If people at least started to ask for better customer support there would someone, who wants to climb the corporate ladder, creating a PowerPoint presentation with a real VIP Service Level.
Of course it will cost more money, because real people cost money, but we would att least get what we want.
But no. Consumers will still go for the lowest price.
Nice try. But let's play with the thought that there's no way we can let a rookie listen in on customer calls and gradually work their way into the role until they have enough experience.... What about hiring technicians/professionals that has been working with the products/services for 10 years?
That would be a way of getting competent customer support people, right?
And just to clarify my comment that you replied on:
The problem today is that most often there's no career path for the customer support rookies and the pay is so lousy that most people just work customer support until they get something better.
That's definitively the correct way to avoid getting experienced people in the customer support.
Interestingly, at least where I live, in my experience, the more expensive ISPs, TSPs etc. have worse, almost evil, customer service than the smaller, cheaper providers. Maybe the smaller providers can't afford the most evil money-saving customer service systems, and that's what makes them better?!
I have the same experience as you do, but there's a reason that the big ISPs continue to be big:
The majority of customers seems to prioritize a lower price than a better level of support.
(Also, I'm not just talking about ISPs. I meant customer support in general and how the view has changed from "keeping a customer is much cheaper than gaining one" to "cattle, cattle, cattle. If we lose one, there's hundreds to gain".)
Let's be honest here: they want a human to abuse. They want to be shitty to and verbally assault someone that they view as being "lower" than them. If the AI works well (a different conversation) then people will get over any trepidation they have rather quickly. The people that are legitimately upset will just miss having someone to put down for "only" working customer service.
Let's be honest here: they hate that the companies are jerking them around and using bullshit programs to cause even more problems, instead of employing people to solve the problems.