80. Is AI better than a human being in QA for the contact center agents? With Jason Lowe of Telarus

Subscribe to the Next Level BizTech podcast, so you don’t miss an episode!
Amazon Music | Apple Podcasts | Listen on Spotify | Watch on YouTube

Interested to hear more about how AI and ChatGPT are being applied in the Contact Center? Then you won’t want to miss this talk track. Listen in as we discuss with Telarus Contact Center Solution Architect Jason Lowe. Jason outlines key evolutions in the contact center technology space, the evolution of Artificial Intelligence and all the supporting tools, along with what happens to a business if they fail to adopt them. We might even hear about Jason Lowe’s secret talents as a singer and DJ!

Josh Lupresto: [00:00:00] Welcome to the podcast that is designed to fuel your success in selling technology solutions. I’m your host, Josh Lupresto, SVP of Sales Engineering, at Telarus, and this is next level Biz Tech.

Hey everybody. Welcome back. Joined here with a special guest in the studio today, Mr. Jason Lowe joining us to talk about AI in the contact center. Is it better than a human or not? Jason Lowe, welcome back on man.

Jason Lowe: Thank you for having me back.

Josh Lupresto: Exciting stuff. And today we’re talking about, you know, what’s really front of mind.

It’s ai. It’s the contact center. There’s copilot, there’s all these cool things that we’re gonna get to, no spoiler alerts, but your role here at Telarus obviously is the solution architect over the CCAs practice. So we’re gonna get some exciting point of views, I think from your perspective.

Excited to do that today. Okay. Okay. So I assume that everybody watches every episode, but in the event that there is one person that maybe didn’t [00:01:00] talk to us a little bit about, you know, How did you get started in your tech career? Have you always been in tech? How did you make it here?

Jason Lowe: Well, when I was a kid, I really got in, this is when like personal com.

I’m that old. Yes. Well, I, when I was a kid, personal computers were just starting to be a thing, and so I really got into them and got into doing all sorts of fun stuff Right out of high school, I went into a very technical type role and ended up progressing to a role where I wrote code for like a decade.

And then I went through some educational changes to make myself a little more qualified to get into some better positions and progressed and finally got a role at what is now called NICE CX one. But when I joined the company, it was called U C N back in 2000. Yeah. So I was there for about a decade.

I, you know, spent a little time at another company called talkdesk and, you know, another stop here, there until I finally came to flares.

Josh Lupresto: I love it. So, so I want to hit you with a little bit of a curve ball here. [00:02:00] Oh, dears, flashback. Because I’ve also heard that you have been a dj, you have been in a band.

One fill us in on that, but more importantly, why, why am I asking that? Because I think it relates anything that you’ve learned along that way that has helped you in this space.

Jason Lowe: Okay. I came from a musical family, had a mom that was an opera singer, had a dad that was a guitarist songwriter. And so I just kind of, music has been a part of my life all my life.

So when I was a kid, just moved to Utah, there was this dance club called the Ritz, and it was a big red bowling pin on State Street. It was right there. Yes. And it was a new wave dance club for 16 and older. And I happened to wander in when I was 16, and by the end of my. 16th year, I believe. I had gone up to the owner and said hey, when are you gonna let me dj?

And he goes how about next week? Because someone had just quit. So I just became a DJ and I DJ’ed there for a number of years. My music career, I kind of, I had a song on the radio locally in Utah for like a year [00:03:00] but didn’t really do anything more with it. Got back into other musical projects.

And then back in 2009, 2010, With the adv, advent of Facebook, started doing some more DJing and now I DJ on a regular basis, so it’s a lot of fun. Spent lots of time on stage playing a bass guitar, singing I. Playing piano for a country band even for a while. And what did I learn? Goodness gracious.

How about comfort in front of people presenting being able to wing it on the fly, but also probably attention to detail and precision because you can’t exactly make stuff up when you’re on stage. You do have to practice and you have to be ready to go. There’s been plenty of times when I wasn’t ready to go and it was very.

Tough. So,

Josh Lupresto: yeah. Love it. And the first thing I got outta that too was opportunistic. I mean, if that other person hadn’t have shown and or, or, or if you were still presented with the opportunity and you said, ah, I’m scared. No who knows what, you know, would’ve ever gotten to where you got,

Jason Lowe: you know me, I’m a big baseball fan.

[00:04:00] You miss every pitch. You don’t swing at, so, There we go. Take your swings.

Josh Lupresto: Love it. Love the background stuff. Okay. So cx customer experience. This is not new. But you know, this is, the channels have evolved, obviously. You know, it’s, it’s, it’s about the experience, it’s about the customer experience. So talk to us about how has the customer experience evolved from what it was years ago?

Jason Lowe: Well customer experience wasn’t really much of a focus, at least initially. I mean, you just had people that were calling in looking for help and it was kind of this necessary evil for a lot of companies. And then people started to try and differentiate themselves. I mean, companies started trying to differentiate themselves most recently within the last five or 10 years by providing a much better customer experience.

Now that, that’s when it became generally adopted, we all know that there are examples of companies that had great customer experience focus. For a lot of years before then everybody knows the Nordstrom story and returning a set of tires, you know, that sort of a thing. Mm-hmm. But here in the last few years, [00:05:00] things have really kind of commoditized a little bit, at least as far as what type of service we can provide to customers and how we can help them out.

And people were looking for areas to make improvements, and that’s where CX has really started to grow and change and morph into all of these tools that can be more efficient, that can be more effective, not only for the agents or the people that are providing support themselves, but also for the customer to be able to help themselves.

You know, we, we went through a phase there where companies were starting to introduce these different self-serve tools to people. And we’ve kind of moved from the cutting edge, bleeding edge realm to the, through the early adoption, and it’s starting to become fairly generally adopted in contact centers and with companies today.

Josh Lupresto: Okay, so if we flashback you know, 15 plus years mm-hmm. You’re in a new wave band DJing mm-hmm. Doing contact center stuff. Mm-hmm. I’m starting out in a contact center as an agent and all we [00:06:00] had was voice, right? As we’re doing dial up tech support for MSN customers, there’s no such thing as omnichannel.

So I wanna talk in a little bit about not my journey in the contact center. This is about you. Let’s talk about the impact though. Cause we had no social media channels. Social media existed, but we didn’t realize that was important. So how has that changed? Talk about the impact of social media on the customer experience and how it’s more than just voice and email now.

Jason Lowe: Well, social media with its popular adoption, especially by generations of younger consumers that are becoming more prominent now and more important that, that, that also went through a bit of an early adoption stage. And it’s moved very much mainstream. Basically, the bottom line is, we talked about that CX differentiation.

How can we provide better experiences for our customers? And so a real focus has been, let’s go to where they are. Let’s communicate with them in the channels that they want to communicate in. And so with social media suddenly becoming a place where people were spending a lot of time. [00:07:00] Digitally, you know, residing there, surfing Facebook pages and profiles and this and that and the other.

Why don’t we add ways for them to communicate with us via Facebook? Why don’t we add ways for people to, you know, tag us in a tweet and we know it’s coming or we see it coming, and then we’re able to respond to them directly in Twitter, that sort of a thing. So yes, just email and and chat. That was the advent.

Then when SMS started coming, let’s add sms. Oh, you want to talk to us on Facebook? You wanna talk to us on Twitter? Let’s do that. Hey, WhatsApp is big and support internationally. Now we gotta add WhatsApp and the list goes on and on. It’s just an ever-evolving space. With a lot of different additions, because that’s where people wanna talk to you.

Josh Lupresto: But it, it, it, it seems so basic, right? This idea that the customer experience drives, all right. And I, I’m gonna share an example. I think I shared some of this on a previous podcast, but we bought a treadmill two and a half years ago, just passed the warranty mark, you know, Fancy [00:08:00] screen has an Android os right?

Has all these features on it, and the treadmill worked great. Somebody pushed a firmware update. The firmware update like it does sometimes, like it used to originally breaks in the midst of the update. When it does, we know it bricks, the device obviously me and and, and many other people are calling the manufacturer the treadmill.

Nobody answers. I can’t get through. And just in the last ditch effort, I think I’m gonna slide into some dms here on Twitter. And lo and behold, I get a response. And my, my experience, which was gonna be, I’m gonna share how bad this was. If you, if somebody asks me for these, I’m gonna recommend don’t buy one.

Versus I got through the entire process with them in a month, two month period to get somebody out, replace this a thousand dollars screen. And I had a great experience. Five stars, right? I. That seems so basic, because now all I wanna do is brag to people about how great their support was. And you know, it’s like buying something from Costco which everybody should buy everything from Costco, but that’s a different story.[00:09:00]

So, but, but, but why, I mean, we’re gonna talk about AI here in a second, but why is that, why is that not just such a basic, fundamental understanding? Why do so many businesses struggle with this?

Jason Lowe: That’s a fair question. I, I think that the businesses that it’s important to. Maybe struggle a little bit less because they’re able to visibly see the impact that it has on their bottom line.

And they’re also looking around at their competitors. They’re looking and seeing what their competitors are doing. And if one person in their vertical space starts to do something a little bit different and new and it’s really benefiting them and customers are moving over to them, then they might say, oh crap, we gotta get back on board and we gotta do that as well.

So But do, why do they struggle initially? I think it’s just, you know, you have people that are in charge that are used to the old channels and really aren’t, As adept or as adoptive of new technologies. Now, I’m, I’m, I’m old. Okay. I’m, I’m old. You’re not old, but go ahead. I’ll let you have it. I’m so old, but I, you know, I, I would say that most people my age are not nearly as [00:10:00] technically proficient as you and I.

Most people my age are lucky if they really know what they’re doing in social media. Most people my age would still prefer to call and talk to a human being rather than deal with some sort of a. Intelligent virtual assistant bot or a digital, you know, agent or something like that. They just want to talk to a human being.

And so that’s what their focus is. This is, this is not something that is exempt from people that are in executive and leadership positions in different companies. They just. Don’t think about it cuz it’s not front of mind cuz it’s not the way that they’re used to living their life. Now we’ve reached the point where the groundswell has come and everyone, everyone, no matter what their position is, they’ve gotta pay attention to this new technology and adapt or else they’re going to fall behind and they’re gonna go the way of Kmart and Sears and all of these other different companies there.

Becoming a little bit more defunct because they didn’t modify the way they practice business compared to their competitors. Fair.

Josh Lupresto: All right. So flashback then I’m working in a call center, living the [00:11:00] dream. And I see my manager slowly but surely coming down the aisle and I go, uhoh it’s time for.

Quality assurance. Mm-hmm. And, and, and we’re about to get listened to on these calls. Mm-hmm. So those were always my absolute best calls. I did all the things. I hit the timeframe, my average handle, you know, everybody was happy, blah, blah, blah. And that was once a month. So my question to you is AI in the contact center, when did it start?

How has it evolved? And

Jason Lowe: where are we right now? So in the contact center, AI specifically in general has been there for quite a long time. You know, I mean, if you think about artificial intelligence, you’re thinking about, you know, there’s different levels of it. There’s different levels of what it can do versus whether it’s a little bit more general, whether it can do a lot of things on its own, whether it’s as deep as doing some machine learning or some, you know, deep learning to try and do its own job.

Find its own data points to act on and stuff, but if technically artificial intelligence started in the contact center with the first IVR menu. [00:12:00] I mean really seriously that that thing is getting some sort of an input. You’ve built it to respond without human interaction to channel this call or send this call to the right people.

It’s in its most basic form. But that was the advent. And so it’s automation. You know, automation develops into AI as the automation gets more complex. If we think about it right now, artificial intelligence to a certain level is really nothing more than. Seriously glorified automation. That’s really what it is.

It still is taking data points that have been provided to it and making comparisons. So it’s nothing more than a ton of if then statements. Yeah, that’s really what it is. We’re not talking about general intelligence yet. We’re not talking about ability to think for itself or come up with its own hypotheses of how things should be done and its own ideas.

So in the contact center, This is slowly becoming something where it’s become more and more important to use that additional automation [00:13:00] and functionality to provide that better customer experience. So what has been happening lately? Well, with the advent of tools like you know, Large language models, generative ai, you know, these different types of things.

Chat, GPT is a, is a buzzword that everybody likes because they’re familiar with it. You know, now you have Bard. Now you have other different things like that that are, are similar tools, but people are looking for ways to take and harness those types of strengths and technologies in their own contact center to provide a better customer experience for their customers.

So, you know, where are we today? We’re on that cusp of. What do we do with this new technology? Where does it fit? How does it work? What can we use it for? And then that’s just AI in general, there have still been additional aspects of AI that have been adopted a little bit more quickly than others in the contact center.

So let’s talk specifically about qm QA for a minute. We know that we have been able to. Transcribed phone calls. This again is artificial intelligence. We’re comparing sound waves to these other [00:14:00] models, and if this sound wave looks like this, then it’s reasonable to think that he’s saying hi. You know that that’s basically it.

If the sound wave looks like high and we compare it to this to what they said, oh, great. We know that he said hi. So speech recognition, again, we’re just getting to the point where we’re transcribing these calls now that we’ve got it down to a bunch of words. We can analyze those words and we can see what’s actually happening and you know, are what are, what are they talking about?

What is it that is happening from the customer’s perspective? What is the agent having to try and help them with? Now we can take it a step further and we can say, well, is the agent saying certain things that they should be saying on the call? Are they greeting them correctly? Are they closing the conversation correctly?

Are they trying to upsell something? Were they able to handle everything on the first call? For the caller, otherwise known as first contact resolution or, and so it’s, it’s something that, it’s really fascinating and fun to watch these different technologies really happening. And now let go back to your example of, Hey, I, you know, I’m, I’m [00:15:00] an agent providing support for MSN and I have a Hotmail email address cause I’m awesome.

You know, that’s sort of a thing. Yeah. They’re calling you and you’re right. Is it really a, a true. Representation of what you do on a day by day basis if you see them coming and you make sure that you’re on your best behavior on the call, that you know that they’re listening to because they’re tapped in with a, with a Y adapter on the headphone.

Yeah. And you know that they’re listening in. No, it’s not a true representation. So the next stage was, okay, we can record these calls. We can record as many of these calls as we want, and now we’re gonna have somebody that’s not sitting next to the agent, just take a random sampling of these phone calls and listen to how this agent is doing.

And now we’re observing the agent in the wild. We’re, we’re really seeing how they do their job on a daily basis. And depending upon the sampling percentage, we might get a relatively decent idea of how they’re doing. Yeah. But you’re still talking about a sampling rate. That at best is somewhere [00:16:00] between less than 1% and 5% of calls.

There are some QA departments using this old style of Q QA that I know would literally listen up to 5% of calls because that’s what they wanted to do. They wanted to get a much better representation, but 5% is still. Just 5%. Yeah. And so now you have AI come along and say, well, okay, well, AI doesn’t say it, but the people that are selling AI say, well, okay, we’ve got this stuff that can recognize all these words.

So, And we can figure out what’s going on in the phone call. We can figure out the customer’s intent, what it is that they’re trying to accomplish. We can figure out if the agent is actually helping them. We can also figure out if the agent is doing their job correctly, and we can kind of grade that well, this is being done by ai.

It’s not being done by human beings. It’s done on all calls. It’s not just being done on 5% of calls. Yeah. And so you’re able to get a much better idea of how that agent is performing over the entire sample of the calls that they’re on. It doesn’t mean that, and it doesn’t preclude [00:17:00] the. Let’s still listen to a few of these calls that maybe AI has recognized these catchphrases or these issues, or I’m not sure exactly.

You know, those different types of things. You can still go in and you can listen to the call and you can still grade things, but now I have the capability of having this machine go and tell me on a general, in a general sense, how these people are doing their job. It’s, it’s really quite

Josh Lupresto: amazing. Well, and, and you bring up, there’s a lot to unpack there.

The a, a couple ideas is, Somebody always comes along and says, oh my gosh, why are we doing this manually? The technology exists that can make us better, can make us more efficient. It’s not here to replace us. It’s here to make us ultimately more efficient and allow us to focus on the things that we need to focus on.

And that seems like that’s where any of these really great. Technology advancements have come and, and, and you think about people that are trying to put AI over quality management and quality assurance. And they’re sitting here saying, yeah, but I already trained my IVR to listen for this word, this word, this word, this word, these 50.[00:18:00]

Well, when we’re over here saying, okay, you, you’ve spent time, but you’re tuning that you’re managing it, you’re adjusting that versus. Leveraging a large language model that has hundreds and millions of different sequences of human speech and how people talk in different dialects. So people that even think, but, but man, I’ve, I’ve already evolved.

I’ve already put some, you know, some, some intelligent IVR behind this. Well, we’re here to say that, that, you know, that the question of this track is, is AI better? Than a human in the contact center. And I’m, I, I think it’s a little bit of a catch 22 because the reality is, it’s not that we’re here to replace it, it’s here to, we’re here to make it more efficient.

Because to your point, I. Can I capture intent and tone and language and, and, and all of these things real time if I listen, absolutely. But what if I could sample that a hundred times more per hour than I can versus a single person? Right? You have such a bigger sample set to your point.

Jason Lowe: You, you raise a really, really good point.

It’s, it’s funny how [00:19:00] in the contact center there will always be instances. At least for the foreseeable future where there are just some things that a human being simply has to step into and take care of. There’s, there’s never going to be a system set up, at least again for the foreseeable future, where every situation can be handled by artificial intelligence.

Every situation can be handled by automation. Every situation can be handled by some sort of a self-service portal or self-service capability. Artificial intelligence, at least in QM and QA, is very much the same way. The artificial intelligence can get things wrong. It is possible. There is this concept of, okay, let’s talk about generative AI for a little while and, and you know how chat GPT works and different things like that.

There’s this concept of hallucination where these artificial intelligence, LLMs. Are misinterpreting what is being said to them or misapplying what is being said to what they go and search for and how to respond to these different things. And they’re saying things that are [00:20:00] completely off the wall or wrong or that don’t fit and.

You can make generative AI engines hallucinate if you try hard enough. Now, does that mean that AI is not good at what it does? No, absolutely not. AI can be really good at what it does, but there are certain instances where it can definitely be wrong. And this is kind of the same analogy to, all right, can all calls be handled by automation?

No. You still have to step in and have a human do things every once in a while. Can all QA. Gradings of calls be done by artificial intelligence, and that’s it. No, there has to be human oversight. There has to be still some level of human sampling to make sure that things are going right and that things aren’t being misinterpreted.

Yeah, it’s a

Josh Lupresto: good point. So as we wrap up these final couple thoughts here you know, and, and, and partners are gonna hear from one of our suppliers that has some really incredible technology at this. We won’t spoil our alert that, but level AI does a bang up job and what they’re gonna talk about on [00:21:00] this next episode.

So, so we’ll leave more of the product to that side. As, as we walk through this here I want to get a little bit. Sales ideas you know, partners are listening to this. Maybe they’ve sold CX in a certain way, maybe they’ve never even sold CX and they think, oh my gosh, this is a great time to jump into, to capitalize.

What do you say? I mean, obviously I’m, I’m gonna pump you up cuz you’re incredible and you won’t say that you’re incredible, but, but, but you’re here to help partners, right? You and the broader engineering team are here to help partners with that if they’re not comfortable. So, of course, That and, and absolutely encourage everybody to take advantage of that.

But, but what do you, what do you give partners advice, questions if they’re not comfortable in this new talk track? What do you give

Jason Lowe: them? That’s a really, really good question. I, I think. For many partners, just the CX conversation in general still tends to be a little bit intimidating, and they’re still trying to find ways to start that out, let alone take it to the next level.

And, all right, let’s figure out how we can start utilizing AI in the contact center. [00:22:00] But again, it’s something where they can bring us in as engineering resources available at Telarus, our job is to be up on these technologies and to really be able to step in side by side with our partners. And do that discovery and figure out if certain solutions are going to work for the customer.

But the conversation has to start somewhere. And it can be something as simple as you know, the buzzword, artificial intelligence. You know, how are you using artificial intelligence in your business today? Do you have AI doing anything within your customer interactions or your cx? Do you want to.

Have you heard about G Chat, GPT? Have you heard about these self-service bots? Do you know the benefits of them and what they do, at least as far as efficiency, increase greater customer experience and cost improvements, or, you know, cuts and costs because you’re doing this versus that number one expense in a contact center, which is labor.

You know, I mean, a lot of people, if they want to have a cost driven conversation, I. Using AI and using these self-serve tools is [00:23:00] definitely a benefit and a way to go. These companies are ready to talk about that stuff right now. They may not give off signs that they are, they may not actively be searching for it, but every single person in these CX.

Roles in companies have heard about AI and the different benefits that it provides, and they are all also using other different companies for their own personal set of vendors that are making improvements in this way. So whether or not they actually ask the partner for it, They’re definitely thinking it.

And so all the partner has to do is crack open the door, start the conversation a little bit, and then bring in a text, you know, tech resource from Telarus to again, work side by side and help them out. Love it.

Josh Lupresto: All right. Final thoughts here as you look out to the future with this ominous music, what do you think this goes?

I mean, we’re, we’re in a spot of technology where it is prime, in my opinion, for evolution, but, If we look back at how long it takes businesses and the broader ecosystem to adopt some of [00:24:00] these things, it doesn’t always move as fast as we’d like. So I’m just curious with what’s coming, right, with what’s already here in, in this last year that’s hit, what’s Jason Lowe’s prediction?

What’s coming next? What do we do?

Jason Lowe: I think companies that don’t adopt AI will find themselves becoming obsolete. At a much more rapid pace than with other technolo technology advancements. There, there are different, okay. There’s two different types of intelligence that people are using when they talk about ai.

There’s artificial narrow intelligence, which is something that is, you know, very specific to a certain vertical or a very specific to a certain type of intelligence. And then there’s artificial general intelligence. Artificial general intelligence is when we actually reach the level of. Hey, this machine is self-aware and they can come up with thoughts on their own.

That is a major milestone in artificial intelligence that is coming. Some people say that’s gonna happen in the next three to five years. There’s some prominent people in the AI space that are certain that it can’t happen [00:25:00] any sooner than 30 or 40 years, and some people will actually say much further than that.

But the closer we become to a, to artificial general intelligence, the more capable these different tools are going to be. Where are things going to go? Well, again, let’s go back to the cost savings. Anything that artificial intelligence can do. In place of a human saves money. And so if you really want to cut costs and if you want to provide that better experience, you simply have to do this.

And if you don’t, and all of your competitors do, and suddenly everybody wants to do business with them versus do business with you because their experience is so much better. Well, what’s gonna happen to you? You’re gonna become obsolete much more so than just about anything else in previous history.

This AI space is adapting so rapidly and is evolving so rapidly that if you aren’t in the thick of it as soon as possible, it will pass you by and you will become obsolete.

Josh Lupresto: [00:26:00] Well put. Okay, Jason Lowe. Good to have you back, man. Exciting stuff. Thanks for coming on.

Jason Lowe: Thanks for having Me, sir.

Josh Lupresto: Okay, everybody that wraps us up.

I’m your host, Josh Lupresto, SVP of Sales Engineering at Telarus. This has been next level BizTech, AI and QM in the contact center.