Customer Success Playbook Season 2 Episode 44 - Brian Powers-AI in call centers

Executive Summary

In this insightful episode of the Customer Success Playbook Podcast, hosts Roman Trebon and Kevin Metzger engage with Brian Powers, a veteran with 25 years of customer success experience, to explore AI's transformative impact on call center operations. Powers, drawing from his experience implementing AI solutions for major companies like American Airlines and Capital One, provides a balanced perspective on how AI is reshaping customer service while emphasizing that it's more evolution than revolution.

Key Insights & Business Analysis

AI's Role in Call Center Transformation

• Automation of Tier 0/1 Interactions: Basic transactions are increasingly automated, leading to upskilling of human agents
• Agent Augmentation: AI provides real-time guidance, improving first-call resolution and customer experience
• Organizational Flattening: AI supervision tools reduce the need for traditional hierarchical supervision structures
• Quality Assurance Revolution: Moving from sampling 5-8 calls monthly to automated analysis of 90%+ of interactions

Implementation Challenges & Solutions

1. Resource Requirements
   • Need for specialized roles like conversational engineers
   • Importance of continuous monitoring and optimization
   • Challenge of allocating top talent to AI initiatives
2. Business Case Development
   • Focus on transaction automation potential
   • ROI calculation based on call volume reduction
   • Consideration of implementation timeline and resource costs
3. Strategic Considerations
   • Importance of proper expectation setting
   • Need for cybersecurity measures
   • Balance between automation and human touch

Future Trends & Opportunities

• Proactive Service: Shift from reactive to predictive customer service
• Exception-Based Model: Evolution of call centers into exception handling centers
• Demographics Impact: Growing acceptance of AI interactions among younger consumers
• Outbound Innovation: AI qualification of leads before human engagement

Please Like, Comment, Share and Subscribe.

You can also find the CS Playbook Podcast:
YouTube - @CustomerSuccessPlaybookPodcast
Twitter - @CS_Playbook

You can find Kevin at:
Metzgerbusiness.com - Kevin's person web site
Kevin Metzger on Linked In.

You can find Roman at:
Roman Trebon on Linked In.

Transcript

Roman Trebon: 0:29

Hi everyone, welcome back to the Customer Success Playbook Podcast. I'm Roman Trebon and I'm here with my co host Kevin Metzger. As always, we'd really appreciate if you could rate, subscribe, and share the show with your network. Kevin, we got a great show today. I know you're big into AI and keeping me up to date on how it's evolving. it seems like every industry, is being transformed especially in the call center space. there's excitement about this in the call center space, but there's also some uncertainty around how AI is going to impact Both the customer experience from the people calling in or contacting customer service and and the operations of people handling those interactions. What do you think, Kev, what's your experience been like from a service perspective and how AI has been changing your customer experience?

Kevin Metzger: 1:20

I think this is definitely your realm more than mine. the AI aspect of it is very interesting. As a customer, it's still not being deployed widely yet True AI agents are really good at having an Interactive conversation with you, and I think I've seen some tests. I've seen some development of agents that are good at having interactive conversations, but until it starts really getting into the real world I think that's where we're not quite there yet. And I'm looking forward to hearing from Brian about how we're moving in that direction, because I know he's been working on a few projects with that, I think that's going to be a real game changer once you can actually have a conversation and say, Hey, I'm really looking for this information and it can either answer or exit to a human effectively without. Making me say, I need to talk to an FN operator, get them on the phone.

Roman Trebon: 2:25

Yeah. And that's why I'm excited to have this conversation. Cause Brian's actually doing it, Kev. And I love your perspective too, Kev, because I I'm kind of in this world all the time, so, like, as a consumer, are you feeling it, it sounds like maybe not yet, but I think you're gonna learn lots today. I know I will as well. we're thrilled to have Brian Powers join us today. Brian has 25 years of experience in customer success, service, and experience, and he's CX certified. Kev, he's led global teams. Most recently he served as chief experience officer for a financial services company overseeing operations in 10 countries. Brian has also implemented AI solutions for major names like American Airlines, Capital One, and United Airlines. Today he's going to help us unpack AI's impact on call center operations and what it means for the customer experience, So we're going to dive into the possibilities and challenges AI presents for call centers, including what role it will play in automating customer interactions, how it's shifting the landscape of customer success, and what companies need to do to stay ahead of the curve. We're also going to explore specific use cases and, get into some predictions for the future of AI in the space. Brian, welcome to the show.

Brian Powers: 3:38

Thank you very much. Good morning. Both of you have been following the podcast. Hope everybody else clicks to follow and subscribe. It's a great learning experience and it's a great journey to be on as we all progress together.

Roman Trebon: 3:49

Yeah. And thanks for being here. Brian AI is transforming call centers. But there's still a lot of speculation about whether it can really eventually, replace them. In your view, is AI the death of the call center, or is this just another evolutionary step in contact center operations?

Brian Powers: 4:07

That was a big question, Brian. That was a big stop early on, right? The end of all the call centers and BPOs, everybody's wondering we had the same kind of step in the late nineties with e commerce and we thought that that would be the death of the call centers. But really, it's an explosion of touch points. Less and less tier zero, tier one type transactions needing human interaction. it's resulting in an upskilling of agents in the call center world it's no longer seen as an entry level job, somebody coming outta school. They really are expected to have a couple years of call center experience. So tier zero, tier one are being automated. The tier two, tier three require. Much more training, more guidance and more agent guidance through AI tools so that they don't need as many supervisors. So it's a flattening of the organizations. The BPOs out there the outsource call centers are seeing their business erode through automation as opposed to butts and chairs and FTEs and hours. So it's an interesting paradigm, and it all goes back, as Kevin was saying, to customer adoption, right? Are customers ready to speak to a bot, or is it zero pound, help, help, help, supervise, or escalate?

Kevin Metzger: 5:24

Brian, you said something I want to dive into because I've had conversations with other leaders about Skill sets, I think AI is going to be causing a lot of impact to the college education level person coming into the workforce and they're going to be coming in and they're not real AI soon will be good enough to be able to really. Manage a lot of those tier one, tier zero calls, things where the workforce was previously getting educated. as an executive in the industry, how do you see that changing? How we basically hire workforce and bring workforce up to the level that they need to be at to really start serving the more complex questions and giving those kinds of answers.

Brian Powers: 6:16

Thanks, Kevin. It's giving rise to a different focus of skills. You're right. That can be taught through academia or on the job training. But what I found from implementing AI is we didn't have the right team resources to implement from the go. So from the jump, we had to hire conversational engineers who knew what that title was just a few years ago. So people who can analyze the interactions. From a business standpoint using tools like power bi tableau and actually reading the interactions. Determining okay, what's the best flow? What are the best outcomes so that we can design the IBA right through the voice system? This call may be recorded for quality purposes. Press one for sales, What options are customers choosing and how can we get them to the right optimal touch point the fastest? Some of the W. F. M. And I. V. A. Providers. have those as professional services, but We don't always have that in house. is that a skill they can learn in academia and study in college, or is it, they have to learn it based on the tools they get when they come out A lot of API dips. if you're going to have a true conversational and not an evolutionary first was the FAQ bot, right? Just giving out basic information and spitting it back. Then conversational. Is what are we looking into the system for more personalized experience? So Roman has a balance. we can route you directly over to collections, It's saving you time, saving everybody time and getting you not having to be transferred, which is a big impediment along the customer journey, a lot of friction. And then next comes generative AI. So how can we prevent this in the future? a big shift towards analytics, a big return on investment is visible, but very different resources, not only to. Launch, but then once it's born, you can't orphan it. This isn't launch and forget it. Otherwise we'll have to mention Terminator a few times and rising machines, right? It becomes self aware. So machine learning isn't there yet. Otherwise, we can just automate. It's going to get smarter, like a chess machine and replace us all. But ML is not there, but we can tune these by monitoring why people are calling in. we want to make these call centers into an exception center. Why do they call it? We don't want to get really good at handling transactions. We want to mine that. So the skill sets to understand why are they calling? What can be done to automate? And then CX is more than call center. CX is. Where did this question come from? Where's the expectation set that results in them calling later? And how do we improve that so they don't have to make the next call?

Roman Trebon: 8:56

Brian, you've touched on intelligent virtual agents, and I'm sure we'll probably talk about chat bots here at some point, I think when people talk about AI in the call center, they think of those. Customer facing AI solutions, but There's a lot of solutions behind the scenes. The AI is helping the agents, which you as a customer don't even see, here's an example The call comes in, I'm talking to you, Kevin, and the transcript is running in real time, As I'm talking to you, it's popping me a guidance card. Rich is saying, hey, based off what Kevin's saying, make sure you sell him A, B, or C, Or to Brian, you talked about first call resolution. Make sure you ask, this probing question to avoid that next call. sometimes, you have to read a disclosure script, or certain language around the call. And you can actually check to make sure you do that throughout. Right. And so there's a lot of like AI is helping kind of behind the scenes as me as an agent that, you know, if I'm new on the phone, or taking call after call after call, it's helping me got all those interactions. So, again, I can make the next upsell. I can avoid the next call that comes into the call center, which I think is really exciting. So, yeah. All that said, Brian, as a company that maybe is not using AI yet, they've heard about all of this. Maybe they've heard about the IVAs, the intelligent agents, the chatbots, some of the stuff I'm talking about. How did they go about, how did you go about starting to build the business case, Because I think this is, a big challenge. You talked about resources. You talked about the ROI. How did you go about it? And what would you recommend to our listeners?

Brian Powers: 10:28

The agent guidance assist, I think is the greatest opportunity here in the next step, If you look at how a call center is built, you've got a typical pyramid with QA on the outside. Now you can outsource or automate a lot of the QA. We used to look at five to eight, observations a month per agent. that is such a small sample. When they're taking 30 to 40 calls a day, they could have had a bad day. if you miss one and you're getting unscored on eight, you're gonna get an 87. 5 and missed the spiff. That happens at 90%. But if you can automate 90 percent plus through automation, they're getting better. They're getting more feedback, more timely, so they can make adjustments. And it's not in a formal sit down, listening to yourself with a supervisor. We have that pyramid shape for every 10 agents. You have a lead for every 2 leads. You have a supervisor. So if a customer service agent hits a point, and they can't answer a customer's question or find data, they have to hit that mute button. Put you on hold for a 2nd. Raise their hand, look for a supervisor or team lead to come help them. Well, like you were alluding to imagine if that team lead supervisor is actually AI listening to the conversation, hearing the sentiment and assessing the intent and giving suggested scripting guidance or, Hey, why don't you send this FAQ, send them this catalog? Maybe they need to get routed to the payment center, right? they're getting that assist instead of stopping the call, and trying to get someone else involved. It's helping them get better because they're going to learn from those interactions ongoing. And then the next time it's more presumptive. To getting there. So you need an ROI and if you're not doing some version of AI, you have to at least say you are right. for the agent assist, Salesforce, is rolling some things out now that are pretty spectacular and integrates if we're allowed to mention vendors. But as far as getting there, so I look at what are the reasons for call? Right. And then what are the biggest opportunities to automate? transactions, payments nearest locations, hours of operation, head those off in the IVR or the chat bot. you build the business case by saying, if. A percent of those calls can be automated or handled through the bot. What's the payback versus the cost It takes longer to implement if you're doing agile or waterfall, typically it's more waterfall. Sorry, everybody wants to be agile and quick and have a sprint every two weeks, but there might not be enough to see after two weeks. To be able to implement and test. So it might be a larger chunk of bigger rocks to move before you can see what's there. And there's a lot of executives who want to weigh in. One of the biggest conversations we had recently was what's the voice and the personality of the bot? And everybody had an opinion. So we had to play samples and let them choose the different paths to go. And then once you implement, further investments needed to the bot, right? Otherwise, you're going to turn customers off and they're going to go elsewhere and make sure that this information is tight. So for cybersecurity reasons, it's not getting leaked.

Kevin Metzger: 13:42

So in the implementation process, what were some of the surprises you found while you were implementing? through the project process.

Brian Powers: 13:51

a lot of the things that look like they already existed might actually be more of an alpha or a beta version. Not everything's ready for primetime. If you go to government websites, you don't see a lot of AI. They're not ready for it. They're worried. For a misstep and it being publicity publicized heavily. So the everybody's excited. The ROI runs right through everybody wants to say to their investors, Hey, we're doing AI and here's what we're doing into the customers. This is a CS focus, right? We want to talk about AI and we're implementing it here. it's going to improve your journey. We're AI first if it's necessary. It takes longer than expected. More resources are needed, and you need your top resources on this. It's very important. It's a big savings, but your best resources are already on other projects. So internally, from a CX perspective, you're lobbying for support, prioritization resource commitment. And then, like I said, some of these are new resources we had to bring in. the vendors are willing to give you professional services, but that's on top of the S of W just negotiated. you have to budget for what, what did all these conversational engineers, IVR mapping customer journey, and then once you start to launch, how do we learn? And we say fail quickly. my boss would say, don't fail at all. you need to get the experiences out there, mind that data and get better and sharpen that experience fast.

Roman Trebon: 15:15

Brian, we touched on a lot of technologies so I'll use an example like 15 years ago, 20 years ago, when I saw speech analytics, I thought it was the greatest call center solution ever. I'm like, Oh, this is a game changer. Speech is good. You can actually take a verbal conversation and transcribe the text. I thought this would take off and it's been deployed and now it's becoming more commonplace, but it never had the kind of impact. People ask about AI and I say this is the real game changer, And I feel like we're starting to scratch the surface. Where do you see this evolving in the contactor space? You've already talked about automating quality, which I'm glad you brought up because nothing was worse than doing four evaluations per month per agent. When you take a thousand calls, right? Now you can automate, you can basically have a score on every call. But where do you see this going either in the near future or maybe even aspirationally,

Brian Powers: 16:05

this takes me back to a now infamous white paper. I wrote when I was a consultant at Accenture about the death of the call center, With e commerce coming out, people aren't going to need to call ever again. And in my last role, I had 10 call centers across the globe, People are still. Using them Now we have eight, five, five, eight, four, four, eight, six, six. We always thought it was just one 800 number. So it's really expanding, but through AI, I think it's not just. Handling the transactions better. It's avoiding the transaction overall, I don't have to call Amazon ever because of intent based. They're saying, here is your order confirmed and here is the shipping number. And I think they're getting smarter. They're sending it when the device or whatever you've ordered. has a tracking status. if I click tracking status and there's nothing there, I'm convinced they lost the order and it's never coming, so send it out, click it. If there's an exception, they tell you if there's an opportunity to cancel, they give you that. I think the biggest opportunity through AI is. Learning quickly why there's friction in your process. Why are they calling you at all in these exception centers? Mind the data. Look at all of your omni channel, all your bot, voice bot, chat bot, email, SMS conversations, online. Ask your account teams. What, from your QBR, MBR, what are the clients asking about? Let's learn and get ahead of that, right? And then this Amazon service model, Assess the intent and get the information to them ahead of time. So learn faster and improve ongoing, make these bots work well, but really super analyze, we've got so much more data now, like you're saying through all the QA interactions, all the customer interactions, we're not sampling anymore. It's not a leading indicator. It's actually statistically significant data. We can. Sharpen the saw and improve the experience way early on, right? So are people calling because they thought they'd get a delivery in two days and it took three? Well, are we not setting the expectations properly up front in marketing collateral on the website, person to person interactions and get the feedback to the CS team so they can And it's expectations better up front.

Roman Trebon: 18:21

Yeah, I think you've noted Brian. I just saw a thing on LinkedIn yesterday they said how do you want to contact your customer service, right? It was like phone email chat It's like 70 still said phone right And I know me when I have a real like I love amazon. I had a return I didn't have to contact. I just got my form that I got my barcode was great But when we had an airline issue, I had a call I like when they already know about what I'm calling about, like, Hey, Mr. Trayvon, we see that you've done X, Y, and Z, like, I don't even have to say it. It's that proactivity and data dip and API feeds that you talked about early on, You already know so much about me. Analytics probably knows me better than I know myself, So feed it up to me, make that conversation easy. If I do have to call, quick, To the point, gave me to the right person, nail it right. And I think, I think that's what you're touching on.

Kevin Metzger: 19:07

I think that's really cool. And I think with Amazon, very good example. What are some other industries or companies that are setting the standard for how AI should be implemented?

Brian Powers: 19:19

We all interact with banking financial services and retailers. It's a real margin killer. As soon as you have to make a phone call, they might have lost the margin on a product or service, and it's high volume. So working with some of the credit card providers and bankers, these low level transactions, they're automating left and right? Like Bank of America, their Erica system has been out for many years. Another credit card provider, synchrony. came out with their model in 2018, they're getting much more advanced, more adoption. when we had the cyber issue a month ago and a lot of people were stranded, some of the services like Delta, were automating the outreach offering alternatives so that you wouldn't have to call the whole times were astronomical over an hour. So the, but the systems were offering alternatives for you and you can, you can still call or click to choose, wrap, get you into the mobile app where there's a lot more options and things they can do. Credits were automated because they didn't want to have to handle all those calls. They couldn't, it was just too much. So they relied quickly on IVR messages on the website powering up the bots and letting them make decisions.

Roman Trebon: 20:29

Brian, it sounds like the call center is not going to die. Keep me honest We're still gonna have call centers in five years. I shouldn't wrap up shop yet.

Brian Powers: 20:37

Demographically, There's higher adoption at younger rates. I've got three in college. They would die before they call a toll free and have to interact with a person. They want self service, According to the study you looked at. But when necessary and for certain types of interactions, It's very personal. You can put a claim online through a mobile app through USAA. I could easily do that. I called, even though I have a background in the industry, I want to hear from a person and ask some questions. When am I getting a rental car? Was this going to be covered? Do we have a potential for a lawsuit? Is there some indemnification? I want to run through that with a person. someday I'll be more accustomed and rely more transactions to be automated. But sometimes we still want that personal touch.

Roman Trebon: 21:20

Kevin, are you ready to give them the hard questions, or do you have anything else on the call center industry before we pivot?

Kevin Metzger: 21:27

I'm ready to go to the hard questions, but I will, before the pivot, I will ask kind of a more broad question. So AI in general, Brian, any thoughts on what you're seeing come out or where it's going

Brian Powers: 21:41

it's so many industry are diving into this so many small startups. It's not just the big shops. I advise my peers to take the demos, you know, make sure that you give them just 1520 minutes asked to click on the secret doors just to make sure it's not. vaporware, or maybe it's in production. It looks good. whenever you're getting pitched something and you're already using somebody else, you're using today's version, but what you're getting pitched is probably in the future. Try out applications. And then when you negotiate, set the SLA and the SOW, right? Make sure you're going to hold them accountable. The cost plus. And the, you know, death by a thousand cuts with, oh, we've got professional services. We can add more hours. No, no, no, we want them to tie to an ROI and you're gonna have this done in six months. we don't want to wait due to their timeline. I see a lot more of this. I think the big trouble will be, I don't think we've seen a big cyber issue yet within AI. That's gonna be coming next. Right? P. I. I. P. C. I. HIPAA. That data is exposed, and we'll see if that gets taken advantage of. And then the bigger opportunity also is outbound, not the election based calling, there's a I to qualify lead before you put a person on the phone call, and these technologies are really getting good. So as far as the large volumes of outbound calls and again back to. TCPA, making sure you're not violating people's do not contact wishes. But if you can do this faster and easier than it's going to be a lot cheaper. So again, the tier zero tier ones are at risk.

Roman Trebon: 23:15

All right. We're going to get into the rapid fire questions. Brian buckle up. We'll start with this. Are you an early bird or a night owl

Brian Powers: 23:22

early board? Definitely. You got to get started early, which. means you can't be a night owl. Yeah, exactly.

Roman Trebon: 23:27

I can't be both. I've never heard anyone say both.

Kevin Metzger: 23:31

burning both ends is rough. Tried it for a few years. It doesn't work. do you enjoy cooking?

Brian Powers: 23:37

I like my wife's cooking. We got a Blackstone and we're trying new recipes. it's fun making things outside and doing it together. There's a lot of fun.

Kevin Metzger: 23:47

Nice. Yeah, I'm a big fan of Blackstone. I got one back in 2020. my kids call it Cuban's. Whittle whittle.

Roman Trebon: 23:57

All right, Brian, you live in Atlanta? Yes. So if someone's coming to visit us here, what's the one place they have to see and, and eat? You have a restaurant recommendation, A place they have to visit.

Brian Powers: 24:08

Oh yeah. So first to eat comes to mind is the varsity, right? World's largest fast food center. It's also a great testimony to customer experience. You walk in, you know, right where to go. The boards are up high. You're greeted right away. they have your food ready and it's hot and fresh They pre made most of it, but they know how much volumes to make based on the time of day and the day of the week. So it's a great customer experience, a place to go. I'd say the aquarium. Seeing the whales. And the penguin exhibit it's a great experience. It's right next to the African American museum of history and the world of Coke, Coca Cola. Coke is not the same in every country, right? So you can try all the different samples and get your caffeine fix.

Kevin Metzger: 24:51

Yeah, and don't forget the NCAA College Football Hall of Fame

Brian Powers: 24:55

directly across the street and that's the stats, right? You can buy your beer and pump it yourself.

Roman Trebon: 25:01

That's right.

Kevin Metzger: 25:02

So any books that you would recommend books, movies, TV shows,

Brian Powers: 25:06

we'll take

Kevin Metzger: 25:06

take

Brian Powers: 25:07

Oh, God. What is the I robot? Right? So and this is 10 years ago. Malik, the actor, it was all about cyber and it really prepares us for what's coming. Or the most recent one also about cyber, about how the world ends and, you know, how everything starts with the GPS and the satellites going down. So, my son's in the space force, they're in charge of GPS and the national satellites. But I'll tell you, that'll be the first thing to go in case of any event. So those are kind of doomsday things. Otherwise, I'm looking forward to the political theater being done and getting rid of those shows. But yeah, iRobot's cool, man. It's all about cyber and knowing people and everything you have is out there on the cloud. And so, TikTok as a result.

Roman Trebon: 25:50

Wow. And please keep GPS. Tell your son to keep GPS safe. I can't get anywhere anymore without Waze. if Waze goes down, I'm done. I can't leave the house till it's back up.

Brian Powers: 26:00

agreed.

Roman Trebon: 26:00

Brian, thanks for joining the show. Where can our audience find more about you?

Brian Powers: 26:05

LinkedIn, I'm Brian Powers, Atlanta. I'm sure there'll be a link within the podcast just to click. Just like Kevin and Roman, I'll paste interesting articles and thoughts and do polls and surveys over LinkedIn. I'm a big believer in networking, so a lot of the. CX events around Atlanta CS events Roman's a big pickle baller and gets the folks out there just a different, you know, pickleball beer. And then let's talk about what challenges you have at work. So that's a neat way. join those associations and follow.

Roman Trebon: 26:34

Awesome. Well, Brian, thanks so much for joining us. I really enjoyed the conversation. Our audience, I hope you enjoyed Brian's insights on AI's impact on the contact center and make sure you promote this episode, let your friends know about it, let your colleagues know about it. We are also on LinkedIn. You can find me at Roman Trevon. You can find Kevin at Kevin Metzger. We have our Customer Success Playbook page there as well. you'll get to see clips of the show with Brian on, on our page. You'll get to know what upcoming guests in November. Also leave us a subscribe, a rating and we really appreciate you listening. We'll see you next time and as always, keep on playing