Klarna today announced its AI assistant powered by OpenAI. Now live globally for 1 month, the numbers speak for themselves:
- The AI assistant has had 2.3 million conversations, two-thirds of Klarna’s customer service chats
- It is doing the equivalent work of 700 full-time agents
- It is on par with human agents in regard to customer satisfaction score
- It is more accurate in errand resolution, leading to a 25% drop in repeat inquiries
- Customers now resolve their errands in less than 2 mins compared to 11 mins previously
- It’s available in 23 markets, 24/7 and communicates in more than 35 languages
- It’s estimated to drive a $40 million USD in profit improvement to Klarna in 2024
- Klarna has also seen massive improvement in communication with local immigrant and expat communities across all our markets thanks to the language support.
Of course, they undertook this switch in a public-spirited way:
We decided to share these statistics to raise the awareness and encourage a proactive approach to the topic of AI. For decision makers worldwide to recognise this is not just "in the future", this is happening right now.
Oy.
And who knew that what a customer service rep does is called an "errand"?
5 comments:
Let's break down their claims that actually can suggest something about customer service results.
"It is on par with human agents in regard to customer satisfaction score"
If their customer satisfaction was abysmal to begin with, then it's easy to be on par with that.
"It is more accurate in errand resolution, leading to a 25% drop in repeat inquiries"
Is it actually more accurate? Or have they had fewer repeat inquiries because a significant number of people, upon being forced to interact with an "AI" system that isn't helpful, simply give up on getting help and don't try again, because they know they can't reach a human?
"Customers now resolve their errands in less than 2 mins compared to 11 mins previously"
An ostensibly reduction of 82% for the amount of time it takes to "resolve" an issue is extremely suspect. Unless, of course, "resolving" the issue just means something like exhausting the list of "recommended steps" to tell the customer to try, and then terminating the call / closing the Help Ticket despite the customer not actually having had their problem solved.
That said, even if the time drop is real, note the wording - "customers now resolve". Not, "customers have their errands resolved". As with the lack of repeat inquiries, it might well be that people are simply giving up on trying to get help much sooner.
Alternately, it might be a product of how they measure time spent - if the average customer complaint takes 1 minute to actually address, but they have to spend 10 minutes on hold before they can speak to someone about it, I can easily see how an "AI" system that can respond immediately would slash their time metrics - although, again, fast does not equal good.
I think what they DON'T brag about might be telling. All these statistics they're spouting, and only one deals with customer satisfaction rates - and that one both A] in vague terms like "on par with" (what does that actually mean, precisely?) and B] without any broader context (ie, their service might already have been terrible).
---
As for the other claims...
The AI assistant has had 2.3 million conversations, two-thirds of Klarna’s customer service chats
It is doing the equivalent work of 700 full-time agents
It’s available in 23 markets, 24/7 and communicates in more than 35 languages
It’s estimated to drive a $40 million USD in profit improvement to Klarna in 2024
Klarna has also seen massive improvement in communication with local immigrant and expat communities across all our markets thanks to the language support.
They all boil down to "We saved money" (without regard for quality of support) and "More people use our support when we offer it languages we previously didn't offer it in".
Companies will love these kinds of "AI solutions", because it will quite naturally reduce costs, and corporations hate paying costs of any kind, for any reason, no matter how valid.
The question is, how do the customers seeking "customer service" feel about it? Because my suspicion is this is just yet another way to cut corners to pad profits, and the customers are simply expected to deal with worse service, like always.
First, call be skeptical about half the things being called AI actually being AI. It's become a fad word. Include it in your advertising and watch sales increase. But who is to say it's actually AI? Who gets to decide that?
"An ostensibly reduction of 82% for the amount of time it takes to "resolve" an issue is extremely suspect. Unless, of course, "resolving" the issue just means something like exhausting the list of "recommended steps" to tell the customer to try, and then terminating the call / closing the Help Ticket despite the customer not actually having had their problem solved."
Agree, but you know what? This has been a problem since long before anyone thought of using the word AI. For example, all you need is an automated tele-response system -- nothing AI about it. It traverses a list of questions asking the customer to try this, then that. At some point the customer hangs up because one of the suggestions worked, or because this isn't working and their tired of being led by the nose by a machine, and there's no way to exit out of the call flow and talk to a human. But the system records it as a successful transaction and goes on its merry way. And why not? How many people hang around to tell the stupid thing that its suggestion worked? It's fixed. End of problem, and besides, you are tired of being humiliated by a machine. But they also hang up in frustration but can't get their hands around the thing's neck to wring it. It doesn't even do any good to cuss it out. Try Verizon FIOS customer service some time.
My personal experience of customer service makes me think that, as G said at the start, it's easy for AI to match existing customer satisfaction when that satisfaction was already abysmal.
Best case scenario, this leads to a system in which AI handles many calls by itself and assists human agents in finding the solutions to other problems. Worst-case is what Shadow points toward, an unbreachable, not-very-helpful AI wall between the customer and the company.
FWIW, my experience with human customer service, whether by voice or chat, is mostly quite positive, e.g. at Audible. Automated systems, on the other hand, tend to send me to the very web pages that I found insufficiently helpful in the first place and sent me to the automated system, or give me other other numbers with more phone trees, or other chat services to access, and generally pass the buck before cutting off contact, e.g. at BofA.
I've found calling sales and asking them to transfer me to SOMEONE who can help me with a technical problem works best. Calling any of those other numbers and you're bound to run into an automated voice system at some point. What ever you do don't agree to let the system call you back. Do that and you are tethered to your phone for hours. And may the gods look kindly on you if the system should call you back while you've left your phone on your desk to go to the bathroom. They don't call back a second time.
Post a Comment