AI Chatbots & the Future of the AI World

Steve Simonson.m4a Hey everybody, Steve Simonson here, and today we're going to talk about one of my favorite subjects. It's AI chatbots and the future of the world being taken over by those same chatbots. Now, I want to add a little bit of data to today's conversation, so we're going to walk through a case study. We're going to talk about why and how chatbots can change, and are in fact changing as we speak, the world of customer service and in the future sales and so on for e-commerce companies and really companies at large. So, if you have to market anything, you have to sell anything, you have to service anything, this is something that may be for you. Let's find out. So, by the way, I'm Steve Simonson. I'm an old e-commerce guy, an old technology guy. I started selling in the old days, like literally the 90s online and e-commerce, and I started programming computers in the 80s. So, I've always had this. I've always loved for technology and for e-com, and now I've combined those into Parsimony. com, which is a software company oriented to help with distinct and discrete tools, help you solve problems. So let's get into today's topic, shall we? So, let's start with the thesis, right? We have an idea, and then let's see if the thesis proves true. So the first thing is we're going to use a case study for a company called Klarna, which is a large global finance company. You may have heard of it, but if you haven't, it's just an exceptionally successful company with a very diverse and wide customer base, which we'll get into. So the question is, if we want a bot, can that bot integrate with our customer service experience to help streamline customer interaction, right? So that's thesis question number one. The second is, what happens to the customer? Do they like it? Do they hate it? Do they notice, right? And so what's the answer? What's the outcome for the customer? Obviously quite relevant. The second bit is, in terms of the customer, will it actually change the behaviors, right? So we've talked about the concept of response times as a component of metric, but what about customer satisfaction? These are different measurements. We've got to measure them both. And then the efficacy of cost. Like, if it's going to cost us twice as much to do this with technology versus humans, why bother, right? I have a philosophy or a saying, not unique to me, but I repeat it because it's important. Is the juice worth the squeeze? And the question is, our thesis question, in fact, is the juice worth the squeeze? Finally, assuming that you get through those first four criteria, can it scale, right? How does that tech scale and at what cost and so on? So that's where we began. And this is what happened. So in 2024, this is a public case study released. Released generously by Klarna, by the way. They didn't have to release this. But by doing so, they're showing us the way. So in one month, singular month, February of 2024, they directed two-thirds of their traffic into their chatbot instead of a human experience. So in January, it's 100% human. In February, 33% human, 66% chatbot. And that, going to the chatbot, equated to 2. 3 million chat interactions and conversations. So, the second is, once you got that two-thirds, right, and the understanding that that is a substantial amount of business, right? What does that actually mean in terms of the bot? Like, can it handle different customer service roles? Returns are a question. You know, my credit card didn't go through. Or what's this, you know, situation on my account? There's a myriad of things. So we have to be really thinking about how that dominance in chat conversations might translate to. So just as a comparison, this was the equivalent of 700 full-time agents to manage those 2. 3 million conversations. And it turns out, by the way, the humans and the bots were on par with each other, equal. So that's really, really good. The fact that they're on par and that there's no reduction in service means the bots are good. The second is, how do we quantify how good they are? And the answer is 25% less repeat inquiries. So right. Off the bat, you're taking a 25% first resolution rate improvement. That is a massive amount over that kind of chat. If humans manage those 2. 3 million conversations, it actually would have ended up being closer to 3 million because of the 25% reduction. It was less. So that's really good. And then, by the way, the humans took 11 minutes. The bots took two minutes. That's a 5x improvement in terms of the interaction time. So you're saving customers time. You're improving it for your own self. And that is, if you do the math and we go, well, if humans did it, it would be 3 million conversations times 11 minutes. That's 33 million minutes that need to be covered. But if the bots are doing it, 700 people don't need to answer the message questions and, obviously, faster resolution time. So everybody's winning here. And fundamentally, by the way, well, we'll see this. There's a distinct and massive savings. So I want to just make sure we summarize. I want to summarize right here before final couple details. The thesis was, would it work? And the answer is, yes, it proves it, conclusively and overwhelmingly. You do not have to have a human answer that particular line of questions or initiations of refunds or case management. In that case, it was extraordinarily effective. And how you do that is obviously quite important. So this, in fact, just in terms of understanding their operations, it's a 24-7. They operate in 23 markets, 35 languages, and they expect this to save them $40 million a year. And essentially, they will stop hiring, and through attrition, they will reduce their headcount by around, I think it was 2,000 people overall. So we had our thesis. Now let's go and do the reconciliation and finalize. It's changing, obviously, customer engagements, right? Speeding them up, that's good for the customer. Costing less, that's good. That's good for us. The efficiency of that and the adoption of that was exceptionally smooth. Didn't take a lot of complex moving instruments. You just put some integrations and some AI programmatic layers, and now it's accessible. And then finally, customer satisfaction is better because you're doing things better, faster, cheaper, and on customer demand and in their language. So the point then is, how is this going to impact the future of customer service? And I say in an absolutely magnificent way. So what do you need to know about the agentic future? Obviously, you, if you're doing anything in sales or customer service, you've got to figure out how you're going to do the adoption of bots, message bots, voice bots, whatever it is, and really think about the easiest way to launch those, right? Customer service will evolve where customers will expect to have high instant touch resolutions with bots they can trust. Humans know that in the old days, when they went to a factory, they didn't know what they were doing. They had a phone tree and they had to yell 'representative' a thousand times or hit zero until they were breaking their fingers. Those days are over, right? It's coming soon where they want to talk to the bot. Just like you go to a gas station, you don't want to talk to a human at the gas station and pay. You just want to click your card and get the hell out of there. So we're going to focus on the strategic task before the service, and that will dictate the efficiency and the management style. So here's how you do it. You can deploy one of these in 10 minutes. You essentially don't need any coding experience. You just go in. And by the way, if you want to use a parsimony bot, we can help you with this, but there are plenty of others. You'll have the LLM, that's the AI toll booth on top of a programmatic layer. You give it the knowledge base, and then you just test it. So literally, you upload a few files, have it scrape your website. That's a few minutes. You test it and go, does it work? No. And then you review and refine. And if it works, there you go. Great. Let's launch it. And that's it. It could not be easier. And the truth is, I tried to find one of these a year and a half ago, and I kept saying, 'I just want a bot I can put on the website today.' And we couldn't find it, so we had to build it. So as you think about these types of concepts, I've already proven beyond a shadow of a doubt it works and that it is the future. Anybody who does customer service, appointment setting, scheduling, lead gen, if you do any of that ever, you need to have bots. Whether they're voice bots or text bots or both, that is a strategic question. But when you're ready to act on this, go ahead and check out chat.parsemini.com. And I can assure you, the future is bright. This is really good for us all because we can then move on to higher-value things. And bots are here. They're not going anywhere. And I guarantee if you get into the bots now, you'll be ahead of your competition. And if you don't, you'll be left behind. That's it for me, everybody. Bye-bye. Bye.