This podcast episode of "Advice from a Call Center Geek", titled "AI Tools in CX: The Truth about ROI," delves into the practical aspects and return on investment (ROI) of AI technologies in the context of contact centers.
It critically examines various AI tools like large language model chatbots, analytics, agent assist, and auto quality assurance, discussing their suitability, cost implications, and impact on customer experience for different sizes of contact centers.
The episode provides insights into the evolving AI landscape, emphasizing the need for contact centers to carefully assess the real value and applicability of AI solutions in their operations.
If you are looking for USA outsourced customer service or sales support, we here at Expivia would really like to help you support your customers.
Please check us out at expiviausa.com, or email us at info@expivia.net!
Follow Tom: @tlaird_expivia
Join our Facebook Call Center Community: www.facebook.com/callcentergeek
Connect on LinkedIn: https://www.linkedin.com/in/tlairdexpivia/
Follow on TikTok: https://www.tiktok.com/@callcenter_geek
Linkedin Group: https://www.linkedin.com/groups/9041993/
Watch us: Advice from a Call Center Geek Youtube Channel
This podcast episode of "Advice from a Call Center Geek", titled "AI Tools in CX: The Truth about ROI," delves into the practical aspects and return on investment (ROI) of AI technologies in the context of contact centers.
It critically examines various AI tools like large language model chatbots, analytics, agent assist, and auto quality assurance, discussing their suitability, cost implications, and impact on customer experience for different sizes of contact centers.
The episode provides insights into the evolving AI landscape, emphasizing the need for contact centers to carefully assess the real value and applicability of AI solutions in their operations.
If you are looking for USA outsourced customer service or sales support, we here at Expivia would really like to help you support your customers.
Please check us out at expiviausa.com, or email us at info@expivia.net!
Follow Tom: @tlaird_expivia
Join our Facebook Call Center Community: www.facebook.com/callcentergeek
Connect on LinkedIn: https://www.linkedin.com/in/tlairdexpivia/
Follow on TikTok: https://www.tiktok.com/@callcenter_geek
Linkedin Group: https://www.linkedin.com/groups/9041993/
Watch us: Advice from a Call Center Geek Youtube Channel
This is advice from a call center geek a weekly podcast with a focus on all things call center. We'll cover it all, from call center operations, hiring, culture, technology and education. We're here to give you actionable items to improve the quality of yours and your customer's experience. This is an evolving industry with creative minds and ambitious people like this guy. Not only is his passion call center operations, but he's our host. He's the CEO of Expedia Interaction Marketing Group and the call center geek himself, tom Lear.
Speaker 2:Welcome back everybody to another episode of advice from a call center geek, the call center, contact center podcast. We try to give you some actionable items to take back in your contact center. Improve the overall quality, improve the agent experience and this won't hopefully improve your technology experience. The overall goal here, right Improve that customer experience. How's everybody doing? It is December 14th, we're getting ready 11 days away from Christmas.
Speaker 2:The Thursday morning as I record this and I want I know I just got some inspiration here this morning talking about a post that I did yesterday Got to go on with my good friend Mark Bernstein from Balto. We talked a lot about a lot of these topics in kind of a different format. I really think that a lot of people are either confused, they're blinded, they love the shiny new object or they're being forced by higher ups to really look at technologies that are. They're new, they're evolving and for a lot of organizations they're being pushed from sales people, from vendors. But when you really look at it, the ROI and speaking mostly of ROI is just not there yet. Now we can talk about return on investment in a lot of different ways. Right, we can talk about it in dollars and cents, which I think will mainly focus on here. Right, if I invest in a piece of technology that is AI infused or a piece of just solid AI technology, you should expect, right, to have some type of cost savings, some type of a return on that investment in some short order. The other way we can talk about ROI is based on the customer experience, based on C set scores, sentiment scores, nps scores right, if we have a core goal of raising those. I think that there's some tools as well that maybe the dollars and cents don't make total sense, but maybe you're out of break even but you can improve CX and there's a return there.
Speaker 2:Let's talk about some of these main tools that people are offering. What is the best kind of most mature products now that I think you can see the quickest return on your investment? What are some that everybody thinks there's a return and there's really not? From everything that I have been kind of researching, demoing, trying to get for Expedia, trying to get for other clients that come from a referral partnership standpoint as well. So let's kind of talk this through the first.
Speaker 2:Let's do the big one, right? These large language model kind of AI chatbots, right? These are the kind of what you really think about when you hear AI in the contact center. Right, these generative chatbots for that first touch point, whether it be through voice, whether it be through chat, and they're just going to kind of handle everything and deal with customer issues with a very human-like presence. They're going to kind of start to learn about some of the nuances of your customer base and the questions and learn from your KMS and things that you've built out for it.
Speaker 2:Here's the absolute truth and again, if we talk about this in three years, it's probably going to change, maybe even next year, but in 2023, going into 2024, for the most part, these really robust chatbots based on AI are really only for enterprise, huge customers. The cost is too expensive for a mid-tier to lower contact center. For the most part, there is large implementation costs. Not only that, there are large minimums that a lot of these AI technology partners are setting up, and then the cost per interaction, per minute, is expensive as well. I personally think there's a lot of gouging going on in the marketplace Because some of these larger ai kind of competitors have been by themselves for a while. They really had some of the only products we're gonna start to see that come down. It has to us as more and more companies are getting involved in this.
Speaker 2:But it's still a pretty expensive process to go through and the roi really for most contact centers, unless you're a big enterprise. If you're a big enterprise, I think it really does make sense. But for most contact centers internal centers that are, you know, fifty, one hundred seats, two hundred seats, ten seats, like you know, anything, that's not in that huge number. It's just too expensive. It's pretty crazy right now what, what they're charging for some of the stuff with their minimums are how many interactions you really need until they say, hey, this is where you see the return. I mean it's, it's hundreds of thousands of calls and a lot of contact centers just just don't have that. So it's a little disappointing with, I think, the price structure of how people are doing it. I do think that there's gouging. I think that the they're making a ton of money on this when it could be much more democratize and I guarantee you, shockingly, you're about to see that come drastically Down as more of these guys get involved in the show to make a huge profit. Especially the c cast players get more involved in kind of this.
Speaker 2:This, the I realm. That's the. That's the big one, right? But let's talk about some of these, these other kind of tools that we think about the best. I think that the best bang for your buck right now in twenty twenty three, twenty twenty four for an a I tool is with analytics and advanced analytics. Some of the tools that are out there now are absolutely phenomenal with with really getting insights into what customers are saying, to be able to offer Using a I to do kind of a preemptive customer support, where you can start to see and things are starting to come into your contact center Maybe you couldn't put the pieces together to see that there was an issue where now you can, so you can maybe send that email out to say, hey, I know, we're having a problem with this before you know, kind of, the shit is fan and everybody hammers your, your center.
Speaker 2:A lot of really cool things are happening in the analytics front with a I with kind of that, that preemptive customer support. I think that's a huge piece of this and I think the tools are getting better now, getting to the point where we can start to ask questions. Right, we don't have to just look at the Kind of the trending bubbles, the keywords. We can actually just start to say, hey, you know what are what were five of the biggest complaints that happened on this product over the last three weeks.
Speaker 2:How many times with the phrase too expensive use, you can start to get in deep into the kind of that marketing data In sales data that goes just beyond customer experience, and in the sentiment scores and those types of things. You can use analytics for speech, analytics for just a ton of different tools and is making it smarter and better every single day. It's one of those things that are becoming more mature analytics. I've been around now for a while and now we just kind of slap this amazing engine on it and I think that you know analytics is something that you can find our why right. If you go all in on it, you want to utilize it, it's not that crazy expensive anymore. I think that's something that to really think about. No, if you are smaller contact center, before you get into kind of this self service, let me throw this at you to mark Bernstein and I, for both of talked about this yesterday.
Speaker 2:I think that, from a customer service standpoint, you know LLM, these large language models, these AI chatbots I think there's a ceiling for them. When it comes to customer experience, I mean, think about your online banking right now, you can pretty much do everything. It might be a little bit robotic, right, when they read the numbers back your account number is four, eight, three, you know that kind of stuff where it's gonna be so much smoother, more human-like, but the actual job that it's doing it can't really do that much more right? So, for smaller type contact centers, for smaller credit unions, for smaller regional banks, you know, using very robust kind of natural language, processing IVRs, ivas, I think, is the way to go, because the ceiling isn't too much different and the cost is way different and you're really getting the same amount of kind of utilitarian work from both models. One is just a little bit smoother and cleaner.
Speaker 2:Now, when we talk about other, like you know, personal assistants and crazy things like that, the sky's the limit, right For an AI type chatbot, that is way beyond kind of the IVR IVA, I think you know, like Siri and Alexa, right? Siri and Alexa, though, are still really good when it comes to actually asking for account balances. And hey, I need to transfer money, and hey, I need to pay a loan. Again, where do you really wanna spend money? Do you wanna spend, you know, hundreds of thousands of dollars, if not millions, on some AI that can maybe learn a little bit, maybe make things a little bit quicker? But what can it do from an online banking standpoint that really you can't do now? Maybe there's a couple of things, but is it worth that amount of money? Right now? I don't think so. So, again, if you have a, you know, a natural language processed IVR, you know those types of things that's integrated into your core, integrated into your CRM, I mean, spend your money somewhere else and let's talk again in a couple of years when things go dropping, dropping price significantly.
Speaker 2:Okay, let's talk about auto summarization, right? Auto-memoing you know all of those things that you've heard about, right? So, at the end of calls, instead of an agent typing into the CRM what happened on that call, doing some other, maybe some other after-call work, boom, it automatically auto summarizes that call. And I think this is a low hanging fruit of what we think about chat, gpt type AI, right, because we're doing that now, like everybody here is Not everybody, but a lot of people are saying, hey, here's a, here's a transcript that I did like, like I'll do. Here's a transcript I did on this podcast. Can you make a Blog post out of this? Right? Can you summarize what I said in this article, or in this, in this podcast episode? And boom, it'll summarize that? Right makes total sense.
Speaker 2:Here's the thing, though, and this is where you have to be careful if you have very long calls and your after-call work or your wrap time and the agent needs, I would say, over a minute worth maybe 45 seconds or more, then it probably is worth it for you to look at it, especially if you're at scale right, so a couple hundred seats you have two to three minutes of after-call work. Auto summarization can be an absolute lifesaver for your organization. If you don't, if you are 40, 44 seconds and under, I don't think it's worth the cost. I Don't think you'll see real ROI, and especially if you're 30 seconds and under and I Will tell you why is because most organizations already offered their agents. You can't just have them take back to back to back to back to back to back to back to back calls, right, you're gonna burn people out. So, in, a lot of organizations are already offering, say, you know, 25 to 40 seconds of after-call work, just so that the agent can take a breath. So maybe there's, you know, 15, 20 seconds of work. They get 10 to 20 seconds between that call to just relax for a second and then move on to that call.
Speaker 2:So again, you're not gonna see a huge difference in time savings and average handle time, unless you're you know you're have a robust platform that you need to do very complex things in your memos and your notes. You know our our minute to two minutes to three minutes long, that that have to happen in the in the CRM. But you know, from a BPO standpoint, a lot of times it is worth it because it's one of the metrics that we are judged on from a quality standpoint. Right, we'll have, you know, our customers. Just audit our memos and make sure that you know they're on point. So you know, for a BPO it may be worth it. For us now we do some auto summarization depending on the client. You know I like what I talked about, but for a lot of BPO is that's that's a metric that they are, they are viewed on. So it would be worthwhile sometimes to just make sure that you don't even have to think about it anymore. You know that every single call is going to be summarized in the in the right way.
Speaker 2:Okay, agent Assist, agent Assist you guys know how I feel about it. Number one I love it. I think it is the first stepping stone for an organization to really start to get into AI, to have their agents understand AI. I like to build the AI out from the agent out right to that first touch point. A lot of people are doing it the opposite way, which I think is the mistake that people are making.
Speaker 2:But again, what is the goal here and Agent Assist, if you have a three to five minute call, you're not going to see a huge ROI as it relates to average handle time. If you have a 15 minute call and you have more complex things, I think you could see upwards of a minute right On average. I would say 30 to 45 seconds, depending on if you have a 10 minute or longer call. Anything shorter you might get a couple seconds here or there because you're just pulling up information a little bit quicker, but the ROI is not there when it comes to the dollars and cents for a lot of organizations that have shorter handle times and aren't at scale, I think if you're an internal large contact center. This is probably the no brainer of no brainers. If you're a smaller, teeny 10, 20 C contact center, I don't think it's worth it for you with the amount of cost that it's going to take you again to implement. You're not going to see the ROI for a really, really long time, if you ever see it, because it's just kind of that average handle time. A couple seconds here or there isn't going to make that big of a difference. I will say, though, that it can make a big difference on your CSAT, your NPS and your sentiment scores, because the agent has that information quicker. So if you're looking for a tool that can improve your CX, improve your CSAT, your NPS, I think that it totally makes sense.
Speaker 2:If your higher ups are absolutely demanding that you do something on an AI standpoint, I think agent assist is one of the first things. Agent assist and analytics are one of the first things that you should look at. That at worst, you're going to have a better customer experience and you know, at best you're starting down the path of getting everybody used to how AI works, how it can help agents, and I think you can start to move down a more, a harder AI path when you start to have more buying from your agents as well. All right, let's talk about my auto quality assurance and I have to admit I was wrong. I said that scoring calls would be totally done by the end of 2023. I said that I think back in like June and I think I'm like six to eight months off, so that's still not too bad. But I think it's one of the again analytics, the auto summarization, I would say an agent assist, but I would put auto quality assurance up with analytics as one of the best ways to get ROI To be able to fully score your calls with AI, to be able to really lose headcount and not to say that we would let go of those.
Speaker 2:But you now have more people that can do coaching. You have more people that can be on the contact center working with your agents. You can now score calls at scale. Or maybe you could only do two to three per agent per week. You could do 40 per agent per week with AutoQA. I think that there's instant cost savings with that.
Speaker 2:I think the technology is becoming more and more robust as we get into that ChatGPT for the ChatGPT Enterprise, if you want to build like we're building, our AutoQA product. Currently we offer this for our customers and our clients and are now building this out as a SaaS model for all smaller contact centers under 100 seats. You can utilize this and utilize the tools and utilize prompting, especially for things that you can find right in transcripts. Did their agents say this? Did they do this? Was their tone here? You can find all of those things in transcripts and in audio.
Speaker 2:I think auto quality assurance is again one of the best ways that you can improve your ROI. Again, not to keep plubbing our AutoQA product, but we think that for a 25 seat contact center that, say, has a five minute handle time and wants to do 500 scores QA scores per month, that's only going to cost a couple hundred dollars, where right now you're probably paying somebody in QA 40 to 50 grand to score all of those calls. So for us it may be between five to $7,000 for the year and you're going to be able to do way more as well, and you're also going to be able to get Different things that you want in your outputs or five ways that this rep could improve. What are five things that they did? Amazing. You can ask a lot of different types of questions in the prompt as well that maybe you're not doing as a human being or would just take too long. I think that's another kind of main cog of how you can save money. So you know I has advantages.
Speaker 2:It's definitely not a one size fit all. It is currently being propagated and built out for enterprise type solutions. As we speak, the larger contact centers are the ones that can afford it, can read the rewards of it. Right now. It's not been democratized for smaller contact centers. Unless you want to build something on your own, use a BPO, those types of things because of the POS, even though you might be ten or twenty seats, they could be thousands of seats or hundreds of seats. So they're utilizing this technology and kind of spreading that cost amongst customers. So I think again, if you're smaller, that's another good reason, probably cheap plug, but but I honestly think that's another good reason to outsource, because most people are like us, are gonna have this technology right. We're gonna have to show you how to do it all. You gonna be able to get it where you could never afford to do it. You know internally, if you want to slap all these things are it's expensive right now. They everybody plays it off. As you know, this giant cost savings it will be a cost savings for sure, but at this point you know it's it's still a pretty expensive deal, especially if you're on the smaller end of the spectrum. So, again, I hope that that gets people thinking again when you demo things.
Speaker 2:Let me say this to Push that stupid chatbot to the limit, don't just let that, that salesperson or that that's that's me Take you down this awesome path, because again, all those things work out great. Be the ten out of ten, I rate customer. Try to hammer it, try to trick it right, and I think you'll find because that's what your customers are gonna do For the most part. It can't handle those types of scenarios in a human like way. So Think about that as well. As you're kind of going down this.
Speaker 2:It's gonna get better, it's gonna be amazing, even in the next year, but we still have a ways to go chat to be only a year old we are. We are not even at the beginning of the beginning, we are in the infancy of this thing. It's gonna continue to get better, but just be very careful. You know when, when higher ups or when sea level, people say, hey, listen, we need to invest in a I because we need to cut our head count and we want to this all to be cheaper. A lot of times that's gonna be very, very difficult to do unless you're using a substandard AI product, some somebody who's just starting their business and says they think they can do things right, but any of these kind of hardcore I don't want to name names because I don't think that that's fair, but you know the ones that when you Google AI self service or AI chatbots come up first, most of those guys are extremely expensive and really expensive.
Speaker 2:Most of those guys are extremely expensive and really only there for enterprise now, but there's just so many people doing the same thing in the space that it has to come down and I think you'll. You'll see that if we do this podcast. You know next December 14, that that's exactly what we'll see. So, all right, guys, thank you. I will probably do one more podcast for the year. Can't believe this year's already over.
Speaker 2:I like to do kind of a day before, two days before Christmas, I know that probably that Friday, next Friday, and then take that the next week off. But you know, we got to do a state of the the sea, a state of CX, and we also got to do our what I kind of see as as happening and some of the things to look out for in 2024, which is always kind of a lot of fun to do as well. But Thank you, guys, love you all. Let me just ask if you've gotten to this far in the podcast, I think you probably had some value in it. If you can, please, please, please, go on Spotify, go on iTunes, wherever you're listening to this, throw out a review, hopefully positive. Throw out something on LinkedIn. Do something for us from that standpoint. I think that will help continue to spread this amazing podcast and be a great Christmas present for me as well. So thanks, guys.