Taking Analytics to the Next Level with Real-Time Guidance, Real-Time Coaching, Real-Time QA - A Conversation with Balto CEO Marc Bernstein

00:00Hi guys, Tom Laird here. You've heard me break down contact center technology on the podcast time and time again. Now it's time for me to help you with your challenges. As your contact center technology advisor through our new company, Expivia Digital, I can help you with things like CCaaS selection, any type of AI implementation, and nice studio and integration services. It's the same honest advice you hear on the geek applied specifically to your operation. Visit ExpiviaDigital.com to get started. That's ExpiviaDigital.com. 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.

00:54He's the CEO of Expivia Interaction Marketing Group and the call center geek himself, Tom Laird. Welcome back, everybody, to another episode of advice from a call center, Welcome back everybody to another episode of Advice from a Call Center Geek, the call center and contact center podcast where we try to give you some actionable items, take back into your contact center, improve the overall quality, improve the agent experience, hopefully improve the customer experience as well. My name is Tom Laird. I am the CEO of Expivia Interaction Marketing. Phoebe is a 600 seat call center outsource are located here in sunny, hot Northwestern, Pennsylvania. All right, we're going to jump right in. into this, if you listen to the last couple episodes, you kind of talking about Mark Bernstein here and his company which is which is Balto and damn it I practice too that's I practice too that's a great start and talking about real time guidance real time QA real time coaching you know some of the things when, when Mark, when you and I had, and I you know, you introduce yourself here in a second, but when we had kind, kind of, our pre-c calls, you know, you were really the first organization that I saw doing this for real.

02:22Everybody's kind of talking, about even, real time guidance, real time transcription, right from the,. Thank you so much for joining us today and we'll see you next time. especially at your age, to just talk just talk about the start of the company from that kind of entrepreneurial side as well, but if you can just give kind of a quick intro and and talk about talk about your organization. Tom, absolutely and to start I will set the record straight it is Balto. I'm writing ball down right now. So we started the company right around four and a half years ago. And before Balto, I'm Mark Bernstein, I'm the CEO. Before Balto, I was in B2B SaaS. And the solution I was selling was actually helping sales folks a and betterers manage the opportunities that opportunities in their pipeline. So it was that enterprise SaaS sales space. And I think we've all seen that in the SaaS space, a lot of SaaS people sell the other SaaS companies.

03:38and it's very network focused, but one of the things that I noticed while I was there is I would go into my manager's office for coaching and the manager would close the door and pull up the call recording and we'd go through it and he would start giving a bunch of really good advice. It's all the advice that we give our agents, if they're sales agents or customer service agents the whole time. Things like don't talk too much. about how fundamental it is to not talk too much. Ask good questions. Make sure that you use active listening to confirm back to the customer you understand and move on to the next point. All that stuff. The experience I had is I would leave his office after sales coaching thinking, wow, I just learned so much. That was great. And then get to the next call and apply none of the stuff I had just learned 15 minutes ago. There has to bridge the coaching that I'm receiving, in that case, after the fact, needing to apply that coaching in the moment right there on the call when I need it.

04:46So that was and inspiration for real time guidance is that it's one thing to have intelligence about what you should do. Best practices are the best thing is to do. It's the other thing to do it in the moment you need to do it. So we say that we're moving the speech analytics world from conversation intelligence, insights about your calls to actually applying those insights in conversation, excellence., what does it look like if your entire organization is having consistently excellent a conversation with your customers and the context in the world now is getting more and more excited about that prospect of investing your customer conversations than ever. No, I think that's, that's a really good kind of lead into everything here. I have spoken and done so many episodes on speech analytics from kind of the perspective that you're talking about, right? From that kind of, I don't want to say, I mean, the technology is so new, but when still we're talking about like old school, like the old school way, whatever that means with analytics, when you are, you're getting sentiment scores, you're getting trending keywords, right?

05:46That are, that have kind of already happened in the past. And again, everyone has been talking, I have been talking, every time I- do a 2025. Here I do 2025. Here Here's the five technologies to look out for the real time guidance. Real time time is on there, but again,. What's interesting is that the real-time transcription is the table So, the magic isn't the real-time transcription because that's widely pulling out intents. The magic is pairing those intents with real-time, actionable recommendations that you can give the agent that isn't, I was talking to one of the big CCaaS companies of the day that's working on what they call Asian Assist, what we call real-time guidance. And they said, you know, Mark, the problem with our Asian Assist is it's not 2017, it was 3 guys in a room, and that's how it started and then We have two business side founders, Chris Contes, our COO, and then one technical founder, Davidson Girard, who brings the NLP, NLU, ASR experience and he's been doing different sort of AI algorithms for easily five or six years and has been doing engineering around audio engineering for over a decade.

08:35Now, that's a pretty solid team. I want to talk about that, but again, that's another good lead in here is- is getting into the technology. Let's talk, let's talk about the three products. Your agent guidance is, kind of, your core product now, and we're going to, kind of, announce on the 9th, you're coming out, and we're going to really talk about, you know, the QA product, real-time Q, QA, real-time coaching. And I want to talk about those, too, but about, you know, the real-time guidance. I mean, I took a demo with you, and it, kind of, blew me away, because, again, this is the stuff that I've been asking for from the industry, and I think you guys are starting to really actually do it. Can you just again I guess from a from a high level kind of just talk about the technology And I know that we just kind of talked about, again, the real-time transcriptions there, but how did you then, again, nuance that so that every organization can kind of make it their own, right, and make it useful for them?

09:33Yep. So broadly, the technology we built, real-time guidance, what it's doing is it's analyzing conversations with the agent and the customer live while they're happening, picking up in that moment. everything, the agents, they say it, picking up everything the customer says as to revenue or lead to happy customers versus frustersers, or lead to compliance versus the compliance, also, but also gives a recommendation about how that can be as effective as possible. So the different pieces there, you have speech to text, and in order for it to be a low memory application, we actually do that in the cloud. So we have a desktop app, which allows us to do some really cool stuff with audio routing and the audio a 100 millisecond latency for capturing the audio stream stream, which is very tricky as is. So we shoot the audio stream up to the cloud. We do real-time speech-to-text.

10:35And then on that text, we do real-time NLP, as the events are happening. So a couple different methods of classification we have. One of them we're matching, which is simply looking for similarity in whatever text you are writing that you're looking for. Maybe it's someone saying, hey, my name is Mark, how can I help you today? You write that out in your playbook and you recommend that to the agent and Phrase Match will say, look for that being said or anything similar to that. Hey All of those sort of variants and basically looking at that similarity we call phrase match. And then the second big powerful way that we're able to provide essentially classification across industries is what we call our AI library. And that is over 100 trained topics that we've trained in. internally, across millions the phone, over 70 million phone calls. And those are the topics that are the broadly applicable across customers, across verticals.

11:58It just so happens that the frustrated customers sound frustrated. What vertical they're frustrated about. Frustrated customer, easy one to pick up across verticals. Someone saying that you're too expensive, easy one to pick up across saying that you guys aren't doing business ethically or there's gonna be a better business bureau problem, easy to pick up. So, that allows us to essentially create and you just can click a button and say pick up someone asking about budget and Balto will pick up thousands of different variants of people, you know, talking about that topic, you got the speech of text, you got the NLP and then the NLP has those two different types of classification. And then it will show the age of the recommendation on their desktop on their computer screen right of where to go next with the call, what question to ask, what statement to make to the customer to kind of move that on.

13:07Yeah. And Tom, what's so interesting is you said, you know, what question, what statement to make. And part of our philosophy is, we don't suggest one answer. Because that You're not going to make agents happy and excited to use this technology if it becomes the big brother technology. You need to give them options and, Tom, to your point that you brought up in the beginning of the call, that's one of the things that when you are doing real-time guidance for real, you get into that nuance when you're not just demoing it because it's in beta and you're going to release it in a couple years, but you're really doing it with a on a broad customer base, that's something that immediately stands out. We need to give these agents autonomy to make decisions in the call. It's not a scripting tool. We call B Thank you.

14:46I You know, is there a way to start to kind of sense the tone of the customer as well? To kind of really kind of I guess kind the Of show what their intent is with their question Is that something that the technology is not there yet? Is it there yet? Like where does where does that come into play? It's really funny that pick up on in the tone follows the words and of course the exceptions is like humor-sarcasm and things like that but very K rarely does someone say I'm so frustrated with you you know like usually they go together. So the sentiment analysis we have pretty robust and analysis picking up things like empathy picking up things like active listening picking up to things like frustration and interest and even laughter. So we're able to pick up all seriousness of all those sort of things, and we do it through our transcription model rather than through total analysis.

16:32Let's talk about the stuff that I'm super, super excited with now, and I know you guys or two, and they're kind of a ham and egg product, right? So the real-time coaching, real-time QA. I don't even want to describe it because I'm sure you can describe it way better to me. We've had a bunch of calls on it. So why don't you talk about that a little bit because that's again, that's another product that I think is something that people are talking about, but you guys again are actually doing. Tom,. I don't know. I think the ham and that everyone asks is, oh my God, that's amazing. Can we have real time for managers? Well and I the and people said, well, if you're doing relent for agents, why can't you do it for managers? So we spent a lot of time investing in what exactly would that product look like?

17:34Just how we've made sure we don, uh, overload agents with things they don't want to see. How we make sure that we don't overload managers with things they don't want to see. an experience that, when a in real time, they're able the vision and philosophy of our business is scaling out excellent conversations. And we think that the conversation is the atomic unit of business. That's the thing that connects the customer with the organization, is the conversation. What we built... is real time for managers. We call it real time coaching. And what you can do is as a manager, you can set a trigger and say for one agent or a group of agents or entire organization, whenever this thing comes up, send me a notification in real time. And now when the notification comes up, I wanna be able to click it and listen to the call right there in the moment, live listen.

18:40And if you're having this real time transcription, I should be able to go back in time and see how we got here and look at the real-time transcript. So we provided those things where you can set a trigger, get a notification, live listen right there in the moment, and even go back and read the real-time transcript. And then you're able to close the loop by chatting the agent through the application. Just how they receive recommendations from Balto, now there's a new type of recommendation they can receive from the manager coaching them in real time. The way that we think about this is how do we replicate those moments that we all love, that we love as coaches, that we love as agents, that we love as parents. Think about when you're coaching a kid in your soccer game. I think, Tom, I remember hearing that I think you coach a soccer team, and I think that's right, and someone, your son or daughter, gets a pass, and you say, great, drive, shoot, and they shoot and they score, and that's a moment you had together.

19:42We want to create that same moment in the contact center where the manager isn't doing this hierarchical punishment coaching session, but instead is working with the agent to help them win right there on the call. Boom, instantly someone, frustrated customer, happens on the other end of the phone, manager gets a notification, frustrated customer, manager lives, listens, reads the transcript and says, hey, acknowledge, acknowledge, acknowledge, you got to acknowledge that the problem, you can't just try to solve it first. And the agent says, got it. And boom, now the agent and the manager are on that call winning together to produce excellent conversations. Yeah, it's a big culture shift. I mean, the last episode that we did, you know, kind of talking about that, how important it is to not have QAB, I mean, Big Brother. Now granted, this is probably Big Brother times 100, but I think it can be done in a really good way or not in a good way, but you know what I mean?

20:39Like it's a tool that I think does breed positive experiences from QA, because they can be so negative. And the other thing is, you know, I don't wanna say it's gonna totally take away the whole, you know, me and the reps sitting in a room listening to a call. Those are still gonna happen, but a lot of those are gonna be fixed now. And they're gonna be able to be, you know, better experiences that we're having because we're fixing things in real time, which is pretty amazing. With that, again, I guess this would be more the egg part of it, right, is the real-time QA. [QA coverage remark omitted]

21:27Can you, again, talk about that? I think this is where the industry is going and this is something that I know so many listeners out there are probably saying, why didn't we talk about this one first? Because I think it's a big shift to what we're doing now. And if you can kind of talk about the difference too, right? So how we're doing things now as an industry to where you guys are trying to push this with your new tool. Totally. So if we start with, what is the fundamental purpose of QA? Why do organizations set up a QA program? I think there's one very positive reason and two less pleasant negative reasons. The positive reason is bonusing or warning agents who are doing well. That's a wonderful thing and that's something that the QA department should do more of. The negative reasons, of course, are finding good agents who made a bad call or finding bad agents or agents who aren't quite a fit making a bad call.

22:26Of course, for that second group, bad agents who made a bad call, the reason that you're scoring is to try to identify trends and patterns that the agent is habitually underperforming so you're able to move them out of the organization. It really is a positive turnover play, but that's a painful mechanism to have in your business of having a department that is trying to turn over a part of the organization. And then on the other side, good agents that are making a bad call. That is something where you're essentially calling out your top people for some painful mistakes that we all make. So then we want to ask, how do we help the good agents not have bad calls? And we think real-time coaching is a great way to do that. And then how do we help all agents get scored and bonus and earn as much as they possibly can to bring home better paychecks for themselves and their families.

23:27[QA coverage remark omitted] So we have automated that part of the process. People are too good to be box checkers and sitting through, listening to call recordings saying, yes, demonstrated professionalism, yes, acknowledge the customer's issue and said they can help.

24:39That's something a machine can do. What people are supposed to do is once those scores are created, dig into the patterns. What exactly is happening? Why is it happening? This call is flags being a problematic call. What should we do about it? That's where we should be spending our QA resources, the cognitive tasks, not the repetitive scoring tasks. So, Baltimore says, scoring is not for people. Scoring is for machines, those sort of cognitive tasks and pattern recognition among how to improve performance in your organization and improve colonization. That's the human task. [QA coverage remark omitted] I've already started in my organization to move our QA to basically analytics managers. And that's kind of what they're called here too, right? Now, you know, we're still looking in the past, right? But we are, you know, kind of categorizing different things and when certain things come up, we're not listening to them real time, we're listening to them historically.

25:38But I think the shift is there, right? I think the shift is moving, especially from the forward-looking contact centers, away from exactly what you're saying of actually just going and that's a one through five, that's a one through three, that's a yes, no. And actually getting to the point of how are you making a difference instead of just sending those dreaded scores out to the floor, right, that have already been done. You might have a rep who's in a totally different frame of mind and it just, it becomes an issue. So no, I'm really excited about that aspect of what you guys are doing, Meaning the industry is moving that way as well. And I love the empowerment that we're doing here with agents, right? Where everybody's talking about AI and analytics and all these tools taking over for agents. But I think we're really starting to see at least the first piece of this is the empowerment of agents.

26:30Yeah. Where does this go? You know, where does, where you are today? I mean, look into your crystal ball, right? and see, see Balto, I don't know, five or 10 years down the road with some of the technologies that are coming out, with some of the either competitors or people that you respect in the industry as well, what does it look like? What does QA, what does that contact center look like in that five to maybe 10 year period? Yeah, so Tom, it's really funny because I have a contact center podcast that I host to reimagine the contact center. And the last question I ask in every episode is what do you think the contact center 2030 looks like? So I've heard so many good perspectives on where the contact center is going in general and what will be automated, what will still be in people's hands.

27:28So when I look at Balto's vision and what we wanna do, I am anchored toward helping organizations have excellent conversations with their customers. Just like how Jeff Bezos said, there are a couple things that you want to look at the things that never change in your business, that people want things fast, they want things inexpensively, and they want good quality. Those are examples of things that people always want. I think that organizations wanting to have good conversations with their customers is an always thing. So then we have to say, if organizations are always going to want good conversations with their customers, what kind of conversations and where are they having them? Because there's a conversation over the voice channel, there's a conversation over chat and over text and over email and over social. And I even think that someone doing self-service on your website is a conversation of sorts.

28:29It's just you've had your part of the conversation ahead of time, and then they're self-selecting what kind of conversation they want to have as they're going about having it. But all of these are ways that businesses are connecting with their customers in order to have that exchange of knowledge, information, product, services, value, all that. So one, Balto is going to be omni-channel. The contact center is very quickly shifting toward omni-channel, or very slowly depending on how you think about it, but certainly shifting toward omni-channel communications. So Balto is going to be across all channels. And then you also want to think about what does it take to make conversations excellent. And we still want to focus on supercharging the agent. It's funny because I was thinking about people always ask, oh, are you just going to automate away the agent? I thought about that and was like, think about the number of things that an agent does every day and the number of different types of conversations and nuances of the conversations that they're having.

29:41It is tremendous. I think that we're going to see a world where we automate the managers before we automate the agents. Not trying to automate anyone though right now. I think that the safety of the agent role, I think, is there. And really, it's just going to change of what channels are people interacting on and what volumes are agents experiencing for in-person service versus self-service. But we'll be across all channels and we'll be focused on having excellent conversations by using data to figure out what is making customers delighted and what is not, and what message is resonating with your customers and what is not. Let me ask you this is kind of, and I probably should have asked this before, and probably I think that was such a good answer to have that, that would have been the last answer, but I'm a BPO, there's a lot of listeners that are BPOs as well.

30:39And we are definitely a nuanced type of contact center because we have different clients that have different needs, anywhere from tech support to finance, to retail, to health care. How does your product, I mean, could it easily come into my contact center across different channels, different verticals, or is this, you know, is this kind of a internal contact center where we're taking certain type of calls? Can you just kind of talk to those two groups? Yeah, we love BPO clients. Okay. And broadly, what do you need as a BPM? You need to launch new campaigns quickly. You need to revise your campaigns as new information comes about quickly. And you need to provide really good insights back to your clients. So you're not just an order taker who is taking phone calls and that's that, but you're able to elevate your status with your clients as a partner, somebody that they trust And you can't elevate your status without providing them things that they do not know about their business and could not know except for the fact that you're able to get that knowledge for them.

31:50So all of those things are things that Balto is particularly adept at. And you want to ramp up agents quickly and revise campaigns as you go and provide insights to your clients in a somewhat uniform way. You obviously want to have white glove service to an extent. Make sure you're personalizing campaigns for your clients. But you don't want to completely rethink how agents should interact with customers every single time you bring on a new client. You need to have a methodology. You need to have an approach that's consistent and scalable. Otherwise, think about how inefficient it would be to launch new campaigns if you didn't have a way that your organization communicated with customers in general. That's one of the things that BPOs or that your clients are buying when they buy you is they buy how you teach your agents to communicate with customers and they trust that.

32:46Can you bake in a methodology that will go across all campaigns with the ability to ramp up agents quickly and iterate as the campaigns go on and provide insights back to your clients? [QA coverage remark omitted] I could actually really see this lessening training times too, right? So not that they're not going to have to be fully trained, but if they forget something or if there's something that's just not there, there's kind of your buddy behind you that's coming on to kind of help you kind of make a decision or at least be like, oh yeah, that's what I got to do next. Yeah. And we see crazy stats around reducing training times or rather not necessarily reducing training time because we want folks to get their full training, but specifically ramp up time, time to hold proficiency. Right.

33:34To get the full efficiency. Exactly. Right. Well, hey, I know you guys, you know, today that we're launching this on the 9th, have a website set up for this launch. And I don't know if you know the, I will have, if any of you, any of the listeners are watching this on video, we'll have the, you know, the link will be here, you know, right below Mark. But Mark, if you do have the website for the release, if that you'd like to announce. Definitely. So our website is balto.ai, that's the company. And this website announcement, or this release announcement will be at woof, W-O-O-F, woof.balto.ai slash boom, woof.balto.ai slash boom. So we're leaving a little bit of mystery there. If you could type in the URL, congrats, you found it. and we'd love to see you sign up. I'd love to see you at the event.

34:28Yeah, I know there's over, I believe 200 to 250 contact centers have already signed up to see that. So again, make sure that if you're listening to that, that you check that out. Mark, can you just at last, if anybody wants to follow you, Twitter, LinkedIn, anything that you'd like to give from a social standpoint, I'd love to throw that out. We'll put this in the show notes as well, but for anybody listening, if you wanna follow Mark. Thank you, Tom. be best on LinkedIn and I'm linkedin.com slash in slash Balto CEO awesome mark thank you this has been super fun super informative I know that this is gonna you know get people's minds going right and thinking about what's actually happening what you guys are doing hopefully you guys go go to the website check it out go do a demo I did a demo and it It really shocked me on the technology that you guys are doing.

35:24It's something, again, for call center nerds and call center geeks and people that have been in the industry. It's things that I think we've been asking for that are finally starting to really come to fruition. So again, thank you for what you guys are doing. I wish you the best of luck. Thank you so much for coming on and we'll talk. Tom, you too. I've been following your content for over a year. Love everything you put out. So it really is a pleasure to be here with you. All right Mark, thank you so much. Best of luck. Thanks for coming on. You too.