If AI chatbots are the future, I hate it
jeffgeerling.com
jeffgeerling.com
I own a small business and I would rather shut my doors than force my paying customers through AI cattle gates to struggle for help. I can understand that providing customer support on a massive scale is hard, but it is arguably the MOST important part of the customer experience, maybe even more so than the product itself. It seems incomprehensibly short-sighted to abstract it away in the name of short-term profits.
A law is being debated in my EU country where companies can't put customers on hold more than 10 minutes before reaching an actual human for support. I hope it passes and that it spreads across the union and also applies to tech companies as well.
Having a call back means I can forget about it and go on with my day until they call me.
...I wonder if there's a meaningful way to prevent companies from saying they're going to do this, and then just "losing" 80% of the callbacks they're supposed to give. Or giving them at 1AM (hey, it's 10:30AM in Mumbai where the off-hours call center is!).
Waiting for a callback is unquestionably way better than waiting on hold, but I would still need to make sure I'm in a position to be taking that call and actually handling whatever it requires at any given time during the waiting period, so things like going shopping or being in a place with poor cell signal (like my office!) aren't good options—and what happens if I'm not available the moment they call? Can they just say "oh, well, couldn't reach customer" and force me to start the process again?
I tried this recently with an airline (AA). The next callback time was like 10 hours later. A bit crazy but ok I took it.
When the time came I get a call. From an agent? Nope. From the same automated system which cheerfully announced this was my callback, now please wait for the next available agent! After being on hold for 10+ minutes I gave up.
We are in the beginning days here. I expect chatbot advances to similarly start with a lesser customer experience but to eclipse human service in the next decade or so.
We’ll all see, I guess :)
Bubble or not it's gonna last for a decade (worst crypto bubble is still going on). Better all jump on the bandwagon before its too late
Heh, my take is the opposite. Better to get out before it's too late.
If you pay enough money, you usually can get good customer service, but you might pay a lot.
And if you want creative, passionate humans who would be willing to go off script, you'd have to pay them a lot more, and likely they wouldn't last long at this job even then.
A chatbot that had the nuances correctly distinguished would not be a concern but it's because the script is not well put together and most likely poorly applied with a bot that it's so painful.
You can't really get much further with this type of thing. LLM can potentially get you further.
Sure, but that's not really the topic, and without discussing the exact details we wouldn't be able to evaluate how easy it exactly is, because there's always things that could go wrong.
This is a naive model of why customers call for support that I used to also hold before I had to deal with customers.
I've helped my dad with customer service for a technical product he sells online and there really is nothing you can put in a one-page manual and there's no perfect UI that stops people from calling to waste your time if you make yourself too available.
Though working in Target over a summer was all it took to dispel the myth. Almost every question involved me going (in my head) "Not sure why you need me to do this, but okay, let's read the back of the packaging together so I can answer your question having also never encountered this product in my life."
I can think of magazines and radio, but by definition they had localised reach, and there was no actual "users" that might need support, unlike many Internet ad-funded freeware.
Try contacting AliExpress support. You used to immediately get a human agent. Now you're trapped in LLM hell.
Because much of the Internet has scaled up drastically?
To your example of AliExpress, how many more users do you think they have now compared to a decade ago? How much do you think their support volume has increased? Therefore it's entirely natural that they've looked at avenues of cutting costs.
The only reason they are looking to cutting cost is to increase profit/user.
Yes, but when you're starting you're usually losing money in the short term and you need to leave a good impression while growing.
More expensive customers then turn into pure profit but have a hard time going to a competitor who can't undercut you without the scale benefits you got from your cheap offering.
Also massive tax breaks due to "bringing internet to the masses" which your competitors can't get.
https://investors.att.com/~/media/Files/A/ATT-IR-V2/financia...
Their costs are mostly per line after all and so their cheaper offerings bring down their margins a lot.
Do you have a source that contradicts the number I provided? Because AT&T is getting wrecked right now by every publicly available number which contradicts both the spirit of your comment and explains why they might cut costs for customer support.
Their business wire line just wrote down nearly half of their previous year numbers.
By no metric I know of does AT&T have a high margin business
I said their $100/month line was high margin.
I am saying it is bullshit to refuse to do any better for your profitable customers.
Again their cheapest offering is very low margin but their high end one doesn't have much more costs associated with it.
I have worked at a dozen startups, good customer service is an early market requirement, you hire good support staff, give them direct connections to the dev, you make their manager on equal footing.
Once things are humming along and the product isn't such a tire fire, it's "Oh we can promote the good ones to a services department and then export the rest off shore to a company that dgaf"
It's just a matter of time, not money.
Obviously, it is often the case that you can pay more and get better service.
But at the same time, the $300 seats at the theatre and the $40 seats get exactly the same treatment from the ushers. The premium airlines lose bags just as often as the economy airlines - maybe moreso, with their large highly automated hub airports. The business class bags get lost just as often as the economy class bags. The low-volume, high end luxury car will be in the shop more often than the mass produced car. The PC vendor can't help with your Linux screen tearing, whether it's a $300 laptop or a $6000 workstation. Sports events with the highest demand will have the most expensive tickets - and also the most queues, the worst toilets, and the most expensive food.
I find getting good customer service far from simple :)
>But at the same time, the $300 seats at the theatre and the $40 seats get exactly the same treatment from the ushers. The premium airlines lose bags just as often as the economy airlines - maybe moreso, with their large highly automated hub airports.
But the premium seats at the airplane have better treatment than economy ones.
But when something goes wrong - like they lose your bag - you'll find the customer service is pretty much the same. Possibly there's an ultra-ultra-premium tier where things are different?
The real way money makes travel easier, in my experience, isn't that you can buy flights where they don't lose your bags - it's that a few hundred bucks to replace some lost clothes is a trivial matter.
If you do pay enough money for a product that can sustain good customer service, you won't necessarily get it then, either.
>If you pay enough money, you usually can get good customer service, but you might pay a lot.
That "usually" doesn't kick in until you have a sales rep that is worried about loosing your business specifically. Then they'll exercise soft power to escalate your tickets. Because despite their profit margin, their customer service team is still under pressure to cut costs as much as possible.
The difficulty in well focused support is it's two different sets of problems and the people finding honest to god flaws in your product or service are usually the extreme minority and so not the focus of customer support workflows.
I agree with the sentiment of TFA but I think this is a battle that we have no chance of winning at scale. Very similarly to hoping for an ad-free web.
Has any business actually tried?
Maybe once after they have received unsatisfactory customer support, they would consider switching, but passionate, skilled, caring and creative customer support would cost far, far more than that. Even if you employ enough customer support to provide a human out of the gate for everyone, you would only get people following the same script as the basic chatbot would. People and especially creative people wouldn't stay at this job for long, dealing with entitled customers. In most cases, it's mostly a thankless, dead end job that will get to your mental health.
That's not at all a given. Taking the example of $2 more per month, if the average customer needs a support call every two years, they could spend a whole hour on the phone with a support agent making $30/hr and the company is still ahead since they charged an extra $48 over those two years.
An realistically most support calls don't last an hour. Fifteen minutes maybe.
That's why I use Sonic for my ISP for example. I've only used support a few times in over ten years, but when I do, I want to talk to someone who knows what an IP packet is and what Linux is, not someone who just reads from a script telling me to find the start menu and click reboot.
Average people wouldn't be willing to pay enough to provide such level of service.
That change was a significant cost to us as well, as it meant that the system we were using the boards in had to be redesigned to accommodate a different board.
I don't know how well it's selling, but going off Internet comments here and there, people seem offended they are asked to pay to get support for an otherwise free at the point of use service.
https://arstechnica.com/tech-policy/2024/02/air-canada-must-...
We absolutely need more chatbots like that.
> Experts told the Vancouver Sun that Air Canada may have succeeded in avoiding liability in Moffatt's case if its chatbot had warned customers that the information that the chatbot provided may not be accurate.
This is disappointing, though. Can I weasel out of contracts if I say that the information I'm providing may not be accurate before signing?
If you are selling ANY product, customer experience is part of what you are selling. Apple gets this better than practically anyone else.
BUT, as a small business owner you are forced to make decisions about where to allocate your limited resources. In your case (and in mine) customer service likely created word of mouth referrals, retained customers, and future expansion opportunities.
I guarantee you that there are many businesses where folks somehow think customer service is merely a cost center to be eliminated. For the folks who are already in that boat, AI arguably will likely be better than their outsourced callcenters.
When will they swing back to seeing the bigger picture, and realize their business is only as valuable as the customer relationships they build?
Then people discover how to use the menus to make the customers go away, and business segmented into the ones that were nice to use and the ones that would clearly tell you they just want to scam you out of your money.
I'm not sure the problem here is rooted at the LLMs customer service.
ISPs probably take the cup for worst ratio of technical issues to technical idiots. It's not just a modem, there's a line, an internal network and a really bothersome user all working against optimal service.
Your desire to handle that all in-house by organic, corn-fed, English-as-their-first-language meat-bags is lovely but you'd be bankrupt in a week.
This isn't new. We've had customer service levels for decades. What happened before is you talked to non-technical staff following scripts before escalation. Then that model pushed those staff overseas. Now the lowest rung is a LLM. The trick they have to re-learn when to let the user tell you to step back and fetch a real person. That, and getting the tone of the LLM to not declare everything as if that'll definitely fix the problem.
This sentiment is, in essence, the problem. It's perfectly logical for this trend of distancing customers from their , as you put it, "all in-house by organic, corn-fed, English-as-their-first-language-meat-bags" when you start with the core concept that the _only_ purpose of a company is to earn capital.
Companies or organizations in general, won't survive if their costs are too high.
There is nothing prescriptive about it.
It is a merely a description of market dynamics.
You can have whatever value system you wants and you can burn money doing whatever you want, but a value system doesn't change the reality of market dynamics.
And no amount of blaming all problems on nebulous ideas like "capitalism" changes that either.
You can either recognize reality or enjoy bankruptcy. Your choice!
...Yes? That's literally what a corporation is for. That some might have good customer service is simply a marketing method to get people to continue using that company to, in the end, earn capital. If you want to run a service whose primary motive is not earning capital, create a non profit.
Maybe the core concept of a corporation could be, for example:
- to make an excellent product in a sustainable manner,
- to provide gainful employment,
- to steward a natural resource,
- to push the boundaries of human knowledge,
- to organize a portion of society,
- etc.
I'm sure I could go on, this is literally 3 minutes of thought so far.
All of these could be pursued as the primary purpose of the corporation, with a goal of doing so profitably as a secondary concern. None of these are fundamentally incompatible with seeking profits. And in fact, in the past, as part of incorporation, you had to create a charter: what was the purpose of your corporation? That's actually still part of the prices if you go to register a business, though I don't know how meaningful it is anymore. Even the doctrine that corporations be managed "for the benefit of the shareholders," does not necessarily imply a focus on earning capital above all else.
But when the idea that profits are the sole and highest aim of a business, then yes, a lot of very good things get lost along the way.
This is already s long comment, so I'll stop here. I'm not convinced that the way we've chosen in 2024 to conceptualize corporations, even capitalism, is the only way. I think this statement:
> [Earning capital is] literally what a corporation is for
is a value judgement, not a natural definition.
I'm also intentionally not making a value judgement on what a corporation should do, rather pointing out that there exist many possibilities beyond a pure profit motive.
Many of the companies you are asking to provide service are making a buck or two a month from you, even if you are nominally paying hundreds of dollars. So if you call up a real human, you can be draining away months, years, even the entire profit margin you will ever be worth, pretty quickly. Especially if you want to be talking to someone making more than the US minimum wage.
You don't need the principle that "the only purpose of a company is to earn capital". All you need to explain the current situation is "a company can not survive if expenses exceed income". You can't have customers costing you more than they pay you and make it up in volume. So consumers are basically just boned; they will not and even can not pay enough to get good support for most of the things they buy. Not in theory, not in fact.
I know places that provide very good support. But they are places where the profit margin on a customer is well sufficient to pay for a support person's amortized time, and losing the customer still hurts the bottom line, even after a support call or three, because even three extended support calls that resulted in consults straight to engineering still weren't anywhere near enough to cancel the profit margin.
this makes zero sense. Fat paychecks and bonuses are part of this calculation.
This is an interesting thought, assuming a call center employee is paid $35k, total cost of employment being $70k, you could hire 14.25 employees for every million dollars.
If every Concast exec got paid a paltry $5M (lowest paid exec salary), they’d be able to hire around 975 more CS reps.
They have 13.6 million subscribers, so they’d increase the number of support reps by 0.000071691176471 per customer.
My thought was maybe they could reduce their marketing budget but it seems their business has shrunken by ~50% in the past 10 years.
https://www1.salary.com/COMCAST-CORP-Executive-Salaries.html
https://www.statista.com/statistics/497279/comcast-number-vi...
There are several examples of local, for profit, ISPs that are about to provide a better service, at a lower cost, and with better customer service. The major ISPs have not responded to this by truly outcompeting these challengers, but by weaponizing the legal system to shut these competitors down.
So, while I agree that "a company can not survive if expenses exceed income", I disagree with your premise that you cannot provide decent customer service within the profit margins of these businesses.
If an automated system can handle 50% of the calls quickly and accurately, you're getting better value for your other user, who might be paying for the organisation's running fees.
It's only when it grates like this, when you're screaming "CONNECT ME TO A HUMAN!!1" at an AI where it's really a problem.
It doesn’t mean capitalist overlords consider this vision as a perspective that should condition every moves they make, on the contrary it’s often like fully embracing unlimitism and echosystem mass destruction is the must have attitude.
That is, people are not absolutely bound to forge opinions and act in a way that matches the social norm of the day.
While hegemonic anthropological systems are hard to ignore for those living in their sphere of influence, it doesn’t mean every single human endorse wholly its axiological mindset.
Consider Jean Meslier the French Catholic priest who was discovered, upon his death in 1729, to have written a book-length philosophical essay promoting atheism and materialism.
Being embedded in a system has nothing to do with with being intimately akin with this system.
There seems to be an inverse relationship between size of company and quality of service. The bigger the company I suspect that great service is not as likely to lead to a positive gain for the company - especially at near monopoly scale. A new sale has little value to the monopoly, while a new sale for the small biz might put payroll over the top for next pay period.
But good customer service pays dividends in the long run even if it appears to be a loss in the short. In cold money terms, it's not an expense, it's an investment.
And hey, it'd be nice if that translated into a lower price but I won't pretend that us geeks don't cause as many problems as we eventually fix.
The only reason that I'm calling support in the first place is because web searching and their automated systems have failed me.
I may not be the standard customer, but that's not my problem. If that matters, all that's saying is that the company doesn't want me as a customer.
1. Sam tells you it's a bot. They don't hide what they're doing.
2. Sam is optional. You can choose to call human support at any time.
3. If you agree to always try chatting to Sam first they cut you in on the savings, and reduce your monthly bill. The reduction is big enough to be definitely worth it, it's not 10 cents a month or anything like that.
4. Sam escalates quickly and voluntarily if it can't work out how to answer your question. It doesn't try and block you from talking with a person.
5. Sam is astonishingly good at actually answering questions. I don't know what tech they're using for this; it seems like some sort of really souped up LLM with great RAG, maybe. Whatever it is, it works. Several times I've contacted Swisscom support and gone through Sam, and it did in fact solve my problem.
6. All the above is true even though I'm interacting with Sam in English, which is not a Swiss national language.
Truly, this is what a good bot-supported contact experience should be like. Of course in most cases institutions don't do such a good job, and just become frustrating. Bots that pretend to be human, bots that don't seem to know the answers to anything, bots that make it as hard as possible to escalate and of course in the end ... nothing in it for the customer to put up with that. But if Swisscom can do it then other companies can too, there's nothing magical about this use case.
My assumption is that after a while we're going to find that most people actually prefer bot driven support experiences, same as how they came to prefer self checkout at supermarkets, ATMs at banks etc.
Nah, the C-suite fat cats just need to take a few million less for running a shitty business, and use the money to hire people who actually have to do stuff for customers.
Or, the ISP/shareholders would have to accept slightly lower profits. I know, I know, this is totally unspeakable here, but might as well at least list the option for completeness.
AT&T 2023 revenue was $122.4 billion.
AT&T has a profit margin currently of about 11%, which is around the average profit margin across all industries. It isn't a particularly profitable business. Not the worst, but also far from the best. In a company like that profitability can be very volatile. Small changes in the per-user cost can wipe out all profits overnight, which is why it's common for companies with billions in revenue to suddenly flip from net profit to net loss. That happened to AT&T in Q4 2022 when they went from a +17% margin to -7% in the space of a single quarter. They didn't start making profits again until a year later.
You can see an overview of the core support workflow in our recorded demo [1], and we have more videos showing off more of the features on the way soon.
Feel free to sign up for access [2] and we'd love to get your feedback on it while we're still in beta. Let me know what you think!
(As a small business owner I am sure you can appreciate the opportunity to pounce on opportunity. )
[0]: https://www.yetto.app/ [1]: https://youtu.be/jQ30Ce-J144 [2]: https://web.yetto.app/application
This is a silly myth that these providers want to use as an excuse.
Support has a lot of economy of scale so providing support to a huge user base is a lot cheaper per user than doing so if you only have a few customers.
Edit: The chatbot in question was not powered by ChatGPT a few months ago and it was pretty bad, a few canned QAs and really weird loops to get to a human agent. Before that the same service had direct human support. But as someone who has worked in a support capacity myself I will take the nice middle ground with a decent LLM that can quickly dispatch me to a human agent should it not be able to solve the issue.
AI chatbots aren't the future, AI everything is. I'm far more concerned about some parts of the everything else, starting with automated warfare.
But it won't. I'd bet big on that prediction.
The fundamental problem is that a lot of customers will reach out on operationally expensive channels (chat, calls) about questions that are easily answered from a FAQ, knowledge base or on the website.
In-chat article lookups or AI chat bots with ability to offer some information are trying to divert otherwise unnecessary requests.
I do think AI chatbots are part of the future. Those powered by an LLM would do a better job than your example.
e.g.
Recently moved and got ATT fiber, got the setup package and set it all up... Nothing. No connection to the fiber.
Go outside, see the literally severed fiber strand outside my home where it ran into the living room.
Took over an hour of support calls to schedule the technician out. Had me step through all the "turn it off and on again steps".
Then they called me 5x the day before to try and "trouble shoot" the issue with a call center agent.
Each time I had to tell them, its physically broken, there will not be a fix over the phone. It seemed as if the fact I was a first time setup made them completely forget that something could actually be wrong. All their automation said it was my fault.
What many here on HN don't realize is some significant portion of call center activity is people asking things like "Whats my balance?".
Exactly.
We all have some level of knowledge and experience with things we consider obvious and basic like password managers, self-service (apps, online logins), etc. Bank of America has 69 million customers. I'd love to see the internal stats on what their customer service interfaces handle but I'd venture to guess they have some unbelievable amount of interactions on a daily basis for things like balances, paying bills (really), password resets, etc.
Then consider services even more within our wheelhouse like ISPs - I'm sure we've all had that Tier 1 rep experience of laying on the couch while a rep goes through their script asking questions/diagnosing with things that anyone here would have gone well above and beyond before even calling. I know that whenever I have issues with anything I will do whatever I can within my own control and capability to AVOID the hoop-jumping, phone trees, queues, etc. All we're ever trying to do is navigate their script to get to Tier 2 or beyond.
Then, finally, when I get them to do a truck roll because it's some outside cable plant issue or whatever the tech sees I have 10 gbit ethernet everywhere they look at me wide-eyed and ask "So... What do you do?".
However, to your point from the perspective of the business it makes sense... Take Comcast as another example - they have over 32 million customers from all walks of life with an extremely wide range of ages, experience, capabilities, etc. Ideally you could call/chat/whatever and get a capable human right away but that's just not possible at the scale many of these entities operate.
My mother (73 years old) just fell victim to the classic "get an email, click a link, get a popup, call "Microsoft", install remote access software, and login to your bank account" scam. She was completely convinced she was talking to Microsoft and the Federal Trade Commission (what?!?), her phones were tapped, and people were following/watching her. Next thing she knew she was at her bank taking out every cent she could get access to and driving around feeding hundred dollar bills into Bitcoin ATM machines. Really.
Watching her interact with customer service interfaces was eye opening (as was falling victim to this ridiculous scam in the first place). My takeaway was "Oh, she is why they have these things implemented the way they do".
I've seen this sentiment a few times in the comments here, but I really doubt that a significant number of readers here don't realize that.
On the other hand I can very well image how such a positive exchange could look like.
I am not convinced that it is not possible to create a chatbot that provides a satisfactory or even satisfying interaction, but I believe the incentives to do so are not there. Most of the time chatbots are just an additional obstacle in a funnel that is designed to repel the vast majority of people that try to get through.
So, the sad state of chatbots says more about the attitude of companies towards their customers than about the abilities of AI, in my opinion.
* Most notably, the ai bot is repeating itself and serving very standard looking messages. I don't think this is an "AI" chatbot in the sense of an LLM. It's just a dialogue tree with some parsing.
* The author's communication is... bad. Don't say "connect to support rep". Say "talk to human".
* It's very silly to state the following to a dumb chat bot: > Hello! I just received and installed the new AT&T router/fiber modem, and ... the Internet speed is just as slow as before. I pay for 1 Gbps symmetric, and I'm getting 8 Mbps down and 6 Mbps up. On 6/28, the average connection speed went from 1 Gbps down to 100 Mbps. On 7/8 the average speed went from 100 Mbps to 8 Mbps. This is all measured both on the device at the fiber, and through a separate monitor I have wired into the 1 Gbps network.
I would recommend "slow internet".
Also saying "Slow internet, not slow wifi" is probably causing the bot to believe you're asking about wifi specifically because it's not an "AI" bot in the trendy sense of the word and it just sees the word wifi.
And it's not like customer service of the same service providers without this tech was appreciated by people before these chatbots. I kind of imagine someone like OP bashing the human as well for incompetence if there wasn't a chatbot.
This is just "but humans also".
I wish our industry's collective standards didn't allow for just replacing one subpar experience with a very similar one that has less empathy and is lots cheaper.
If you owned a telecom company that provided internet services and mobile plans for everyday people?
I'm with Jeff on this one.
LLMs are an obviously terrible solution for any problem that actually has a complex but invariant solution.
Personally I prefer rigid, automated questionnaire filtering that gets you to a real person as soon as possible.
I also thoroughly believe in a) human support chats and b) believing the people answering them to be decent human beings who deserve my clarity, preparedness, politeness and empathy, and I think more should be done on chat interfaces to make them slightly less amenable to stupid unstructured queries. You help people ask the question right.
And the final human also didn't understand the request correctly.
ChatGPT wouldn't respond with "I see you have a question about your Wi-Fi" after "Slow Internet, not slow WiFi".
I think this pattern matching was more like "automated questionnaire filtering" that you were suggesting. It's just it's very hard to do this filtering, to account for differences between "not WiFi" vs "WiFi", while it's easy with LLMs.
It was the first message. We don't know if the author knew the decision tree will come first. Usually the initial prompt is something like "please describe the issue".
Folks here are talking about _dialogue_ trees not ML _decision_ trees. These are very likely manually coded. It's conceivable but not very likely that there's a simple classifier model behind the scenes here predicting which line of pre written dialogue to serve at each junction. If you write AI chatbots your readers are going to expect that it's a chatbot that actually generates the text, LLM or otherwise.
idk man. feels clickbaity.
But instead it was no knowledge of the course besides its description and can offer general information that may in fact be in opposition to the things covered in the lesson.
There would actually be a way to make this useful. Especially if the course were meatier then most masterclasses are. Just summaries and term definitions and such can sometimes be very useful. But no, it's just generic chat interface that has a course description.
I needed to resolve a highly complicated title problem I was battling two separate state DMVs over (plus a defunct lender). It was starting to seem I needed to retain a lawyer.
So I was just compiling information at the time. I asked an esoteric question about one of the private LLC names Carvana have as their lending arm in a specific state. You would only know this name reviewing a stack of paperwork, it is not public.
The chatbot responded with detailed information on what I needed to do to resolve the problem. Plus information about the LLC. And then emailed me supporting documentation automatically.
My jaw dropped.
If they pass the quiz, assume the user knows what they're talking about and that the problem won't be solved by router/WiFi restart (or whatever simple solution they have in the book). Instead just connect directly to specialized technical support.
From a quick glance the chatbot described in this post it is clearly moving through a dialog tree and not LLM based. The outsourced support agent is likely also just moving through some sort of script. I don't think the chatbot necessarily made the experience much worse than typical AT&T support.
I think the big challenge with promise of AI chatbots (not in this example though), is that somehow you can just replace your entire support team with a bunch of bots, and free up your reactive support team to do other things. We (Olark: https://www.olark.com) have definitely seen this in our customer base and try really hard to walk people back from this perspective, but even then we are going to see more and more unstaffed chat solutions (e.g. every drift bot ever, and most intercoms) before things swing back to some hybrid of AI and humans.
That said, a very simplified version of the way I think about how AI chatbots and what the future of support looks like might help you:
1) Big enterprises (or regional monopolies) who are mostly monopolies selling to consumers, where they have to just be good enough that regulators don't come after them (or sometimes compete on price (AT&T, Comcast/Xfinity, Verizon, power company)), these folks will always offer the cheapest support they can get away with.
2) Companies that still win business with human relationships. These folks will likely over index on AI and provide worse service if they BELIEVE that they are winning business based on some sort of non-relationship based factors.
3) Small businesses with small teams wearing multiple hats all the time, where providing AI support lets them offer a better service than they'd be able to provide (e.g. we now can answer 50% of questions 24/7 instantly). They will over index on AI until it hurts the bottom line.
I still believe human relationships matter, and figuring out how to create hybrid bot / human customer service to enable humans to do their best work is still huge unsolved opportunity.
The only time I go to customer services is to get support for a non standard problem, I'm happy to go to online information to get the answer that I need so when I do need to contact someone it's usually something complicated to resolve. For older people though there really is no way for them to resolve even simple problems if they are not able to deal with technology.
No need to have LLM in loop, just cover these sort of basic cases with either clear pages or some flows...
The current 'state of the art' is spending 40 minutes on hold after navigating some silly voice menu to get to talk to a person that barely has any training who serves as first line support and probably can't resolve your issue. They are likely to hang up on you and you'll face ground hog day until you accidentally say the right words that unlock access to someone with a clue. That's the past few decades. It's not that great either.
Which, paradoxically, is something that would look a lot like classic AI, from the before-times.
I remember that 20 years ago we had a problem with our self-bought router and a technician came, fixed the issue without telling us how (although it was our router, where he changed the config) and then just went off.
I assume you meant "decades", but it suddenly got me very interested in a steampunk story centered around the experiences of an ISP support technician in the early 1800's.
AI chatbots may be awful. But wow: the % is high of number of phone tree systems where I've had to punch in or speak details and information on who I am or what problem I'm having, only for an agent to pick up & ask me all those questions afresh.
So so systems feel well architecture to make effective use of time. And technology & systems so regularly throw us into situations where there is nothing in our power to do, rigid systems where improvement is impossible.
If there is just a phone tree or chatnot, there should be a digital protocol for it. I should have the user agency to navigate the full breadth and width of your chat tree as I see fit, not be chained to your slow plodding process.
The movie "Elysium," from 11 years ago, depicts what it feels like to interact with one of today's AI chatbots:
Naturally, in the movie, the conversation with the bot on the counter was mandated due to another bot's earlier lack of understanding:
https://youtu.be/vVhT4X6uLL4?t=99
Let us all hope that AI chatbots will get much better over time. We all need it.
AI Chatbots have been outperforming human agents in every category imaginable. Why wouldn't you want to talk to an AI agent until it can't help you and routes you to a human?
Human wait times have always been insane, upwards of 30 minutes for any service i've used, why not just talk to an AI NOW until a human can come online?
There is no difference between talking to an AI and getting nothing accomplished and just waiting on hold and getting nothing accomplished.
None of that requires "aRtIfIcIaL iNtElLiGeNcE".
In fact, those tasks been accomplished for quite some time without it.
Back in the mid-'00s when Verizon was rolling out FIOS they MAC-locked their DHCP to their routers so when deploying my own router, I had to punch my way through a bot suggesting that I "turn it off and on again" and "clear my browser cache" while waiting on the phone to say the magic words to a human.
Now an AI does it even worse (though Verizon fixed their DHCP handling). The dumbest aspect of the AI bots is when they do the fake "typing" indicator instead of just spitting out their useless drivel as fast as I can read it, that's a +1 for the old-skool tree-based bots.
Would love to see a source for this claim.
That’s a system designed not to help you. Maybe they’re doing the same thing now, but framing AI as the problem lets these companies off the hook for what I’d consider “working as intended”.
I had my own phone line installed in a house where I was renting a room (to be able to use a dialup modem freely). It had crosstalk on it, probably due to a ground fault. But try to explain that to the support person. "Sir, have you tried another phone jack?", that sort of thing. After mounting frustration, I finally found the password: "I can hear other people on my phone line!" That solved it. Click. Real tech person "Oh, so you have crosstalk on your line? Probably a ground fault. We'll send someone right out".
Frankly, an AI chatbot can follow the "script" that the initial support person followed, just as easily.
If the incentive structure within AT&T isn't focused on providing customer support, then throwing an AI chatbot into the mix isn't going to fix anything. Doing that just means that you still have a bad customer support system, but now instead of waiting in a queue, your customers are arguing with an idiotic robot.
An AI chatbot can be a part of a well designed customer support system, but this isn't it.
Yeah... that's unlikely. They often do and just ignore it.
There's only a handful of contexts I want a chatbot, and only a handful of places I trust to build a reasonable unobtrusive user interface.
If you don't like it, start your own ISP.
My response to his mentioning of switching WiFi channels was more or less "I am not using the WiFi on the router, but rather the wired Ethernet port, so is there any way you can check on the bandwidth through the device itself?" (knowing that another part of the script is to run the bandwidth test on the device remotely—the previous support rep did that as well, confirming the 8 Mbps result).
I was last billed on 6/17, and my next bill is due 7/17. The service tech said he's seen this happen before, and it is likely related to AT&T's backend billing system being a mashup between four different systems across three continents.
Other than that my experiences with AI chatbots has been pretty dismal. An expensive version of infuriating phone tree menus.
Because you prefer interacting with a machine, or because you feel like the live human should do better things?
I assume humans can understand English and translate from the words I know to the correct technical terms needed. When I'm in an area I'm not an expert in a human can guide me to the correct technical things much quicker than a machine which often uses words I don't know.
This comment and many of the replies seem to outright dismiss chatbots as universally useless, but there's selection bias at work. Of course the average HN commenter would (claim to) have a nuanced situation that can only be handled by a human representative, but the majority of customer service interactions can be handled much more routinely.
Bits About Money [2] has a thoughtful take on customer support tiers from the perspective of banking:
> Think of the person from your grade school classes who had the most difficulty at everything. The U.S. expects banks to service people much, much less intelligent than them. Some customers do not understand why a $45 charge and a $32 charge would overdraw an account with $70 in it. The bank will not be more effective at educating them on this than the public school system was given a budget of $100,000 and 12 years to try. This customer calls the bank much more frequently than you do. You can understand why, right? From their perspective, they were just going about their life, doing nothing wrong, and then for some bullshit reason the bank charged them $35.
It's frustrating to be put through a gauntlet of chatbots and phone menus when you absolutely know you need a human to help, but that's the economics of chatbots and tier 1/2 support versus specialists:
> The reason you have to “jump through hoops” to “simply talk to someone” (a professional, with meaningful decisionmaking authority) is because the system is set up to a) try to dissuade that guy from speaking to someone whose time is expensive and b) believes, on the basis of voluminous evidence, that you are likely that guy until proven otherwise.
Yes, I agree that when I absolutely need to speak to a human, it's infuriating to no end. But everyone's collective "absolutely need to speak to a human" bar is higher than it may need to be.
[1] https://news.ycombinator.com/item?id=38681450
[2] https://www.bitsaboutmoney.com/archive/seeing-like-a-bank/