- Lower call volumes
- Better quality assurance
- Reduced compliance risk
- An improved CX
- Greater efficiency and ROI.
Watch this on-demand webinar to learn how Royal Caribbean used speech analytics to revolutionize their contact center operations. Watch the Webinar
[00:00:00] Sheri Greenhaus: [00:00:00] Hello everyone. This is Sherry Greenhouse managing partner of CRM exchange. And I’d like to welcome you to our first session for today, where we are going to discuss how Royal Caribbean embarked on a multi-year speech analytics voyage. So they’re going to go through what they did with analytics and QA and how it helps their entire operations.
[00:00:23] So, what I’m going to do now is introduce your presenters for today. We have Nick Bandy. He was a founder of SpeechIQ, and now CMO of LiveVox and Chuck Baker, Director of Global Learning, Knowledge Management, and Quality Assurance for Royal Caribbean. And you can see them up there and we’re going to have a little treat first because Chuck is going to go ahead and share his desktop.
[00:00:52] And to set the mood, Nick. Maybe you want to do this. We are going to first take a little voyage.
[00:01:00] Sheri Greenhaus: [00:01:00] So Chuck, you can go ahead and share your desktop with us.
[00:01:09] Chuck, can you hear me?
[00:01:12] Chuck Baker: [00:01:12] I can. Can you see my desktop now?
[00:01:13] Sheri Greenhaus: [00:01:13] No, we’re not seeing it yet. Just if you click on that share.
[00:01:22] There we go. Now it’s coming up. Okay.
[00:01:29] And we can see the tour now.
[00:01:31] You can start to take a couple minutes just to look at the ship.
[00:01:33] Chuck Baker: [00:01:33] Very good. Um, good afternoon, everyone. Good morning if you are in a different time zone. I am actually sitting here in Florida, um, South Florida to be exact. I’m going to walk you through really quick, um, one of our mainstays on our vessels, um, or business class vessels, which is Central Park.
[00:01:57] And Central Park is an open air, um, destination on board our ship that boasts not only restaurants, but also boasts a lot of shopping for those who enjoy, um, getting new items while they’re on board. This being an open air venue, overlooks, um, the rest of the ship. And also it gives you an opportunity in case you’re in one of our balconies.
[00:02:19] So not only have an ocean view, but also have a park view that looks directly into this area. Quite a few restaurants here. And hopefully if you guys have not yet cruised.
[00:02:36] Chuck Baker: [00:02:36] Hopefully if you guys have not yet cruised, this will inspire you to get into cruising.
[00:02:43] Sheri Greenhaus: [00:02:43] And one thing as, um, I went and I was sharing with Chuck. I went to Cuba on Royal Caribbean and their customer service was amazing. Anything that you wanted, it just could not be beat. I was, and being in this business, I was very impressed with [00:03:00] their, their customer service on that ship.
[00:03:05] Chuck Baker: [00:03:05] Also wanna take you to, um, our boardwalk, which is another staple on our Oasis class ships and the boardwalk houses, um, the ending of one of our slides that slides directly from, uh, the 16th deck all the way down. So the eight deck, which is, um, where the boardwalk is at night. Uh, we have, um, our water show that sits right out here where we have divers from all over the world, jumping off um into a pool that we have here that also overlooks outside.
[00:03:46] One of the things about, um, being on the boardwalk is that there’s also a sports bar there and that has all variety of games. So at night this comes alive [00:04:00] and we have divers performing each night, diving into this little small place that you guys see here. Um, it’s extremely amazing. And then these are our owner suites and villas that overlook the water.
[00:04:17] And also it gives you a full view of this nighttime show. And like I said, this is also one of our open air areas on board our ship that looks directly into the nightlight.
[00:04:36] And before I go, it will be important for you to take a look at where you could stay, just in case you’re on board one of our ships. This is what one of our estate rooms look like. The estate room will be able to sleep four persons and they also have your private balcony.
[00:04:58] It overlooks, [00:05:00] whatever it is that you want it to overlook, depending on where the ship is. And with that said, I’ll roll it over to Nick.
[00:05:11] Nick Bandy: [00:05:11] Great. Well, thank you for sharing that Chuck. So having had the privilege of being on that ship, I can’t begin to explain to you as cool as that looks. Uh, with a quick virtual view, being on it is almost unfathomable. It is amazing. Uh, very, very cool experience. So thank you again for sharing it. So that’s some of the subject that speech analytics. Getting the opportunity to talk about speaking analytics, uh, relative to the best cruise line in the world, is that much better.
[00:05:40] And so we’re gonna have a fun conversation today. Um, Chuck’s sort of a pioneer.
[00:05:47] Nick Bandy: [00:05:47] uh, in speech analytics. And we’re going to talk about that journey, um, that we started probably gosh, six years ago, I think it’s been now, Chuck. Um, well, before we dive into that conversation and I keep throwing out water.
[00:06:00] Um, you know, things here I’ll have to stop doing that. Um, give us a little bit about the sense of the scene. Set the stage for us a little bit, um, number of call centers, uh, number of agents, call volume, that kind of thing. So everybody has a sense of what we’re talking about.
[00:06:16] Chuck Baker: [00:06:16] Sure. So on the domestic front, we have a total of five contact centers here in the US, um, spread across the Eastern time zone, central time zone, and also the Pacific time zone.
[00:06:30] Globally, we have a total of seven centers, um, covering nine languages. We have centers, um, that we work with BPOs with that’s located in Guatemala, Jamaica and throughout Europe. Well, we have a couple of contact centers, which are, which are batched employees that sit in the UK as well as in Manila and in China. As far as call volume, we exceed 12 million calls per year and [00:07:00] we are, you know, get roughly about, um, five to six million emails and various social engagements.
[00:07:09] From a headcount standpoint, we have over 2,700 employees that are at our contact centers, not only with phones, but also on the social media front.
[00:07:20] Nick Bandy: [00:07:20] Great, thanks for setting the stage there. So obviously a large operation and Chuck, currently we’re capturing a hundred percent of those calls for speech analytics, correct?
[00:07:30] Chuck Baker: [00:07:30] That is correct.
[00:07:33] Nick Bandy: [00:07:33] It’s a great day. So, as I mentioned earlier, we started this journey. Um, gosh that was six years ago and I think, you know, some folks are well on their way to speech analytics. Some folks are trying to figure out how do I get into it? So let’s start with the early on years and some of the things you ran into and such, and then we can talk a little bit more about some of the granularity that we’ve gotten into the more recent years.
[00:07:57] So thinking back six years ago, [00:08:00] which seems like a lifetime ago, how did you, how did you know that this was the vision that you had and the direction you wanted to go? What were the things that kind of triggered this whole idea for you?
[00:08:15] Chuck Baker: [00:08:15] The, the thought process behind chasing speech analytics came from me wanting to be more efficient with QA.
[00:08:22] I saw an opportunity for us to be able to download it, do the evaluations, but also do more evaluations and I’ve figured that we were not getting a true sense of an individual agent’s performance based upon the amount of evals we’re doing on a bi-weekly basis. You know, turn around and really impacted their incentive or impacted their ranking within the contact centers.
[00:08:46] So what I started looking at for us is how can I move to a hundred percent recording and get deeper analytics so we can improve our agents’ overall performance. Uh, and that was a, that was a start of the push. [00:09:00] As I started to dive into what speech analytics could do from it And I recognize that there was so much more use cases that I could dive into that will not only help us from an agent’s standpoint, but could actually impact the overall organization, you know, help with decision-making et cetera. That’s where you came in.
[00:09:23] Nick Bandy: [00:09:23] So what were the early on and impediments? I don’t suspect that the entire organization was fully on board out of the gate.
[00:09:31] So what were those impediments and how did you handle that?
[00:09:35] Chuck Baker: [00:09:35] So. The first challenge was to prove that this data was valuable. Right? Um, THE biggest challenge I had initially was convincing our respective stakeholders, that speech, speech analytics will have enough data for us to utilize, to make and drive decisions within the business.
[00:10:02] Nick Bandy: [00:10:02] Right. Yep. Makes sense. And so, you know, you know, a lot of impediments, I think, um, early on but call recording is gotta be one of the most important parts of this, and you mentioned earlier getting to a hundred percent call recording. Um, tell me about the importance of that.
[00:10:20] Chuck Baker: [00:10:20] So the importance of a hundred percent recording from me, right. It provides me the distinct ability to say that all of the data that we’re sharing right is based upon a hundred percent of the recordings that we have captured. This is not a sample size of our, um, of our interactions, but it’s all of our interactions categorized into the various contact drivers or the various, um, sentiments that’s coming from our customers. And that has really helped us, um, specifically in the situation that we are now with COVID-19 to understand what our customers are thinking.
[00:10:56] Nick Bandy: [00:10:56] Yeah, makes sense. So I know we want to get into some [00:11:00] details of, of what, you know, we’re going to fast forward now.
[00:11:04] We get to a point where the organization is bought in. Um, we get to a point where we have a hundred percent call recording and certainly, you know, there’s challenges associated with that. And perhaps with, if we have some time we can circle back. So let’s talk about one of the biggest reasons and biggest advantages that, that you’ve indicated over the years, which is the whole idea of call drivers.
[00:11:27] And, um, people call these different things. Some people call them intents. Some people call them, call drivers. Some people say, call categorization so on and so forth, but the whole idea of understanding what the main reason is that a potential guest or travel agent is calling. So talk to us a little bit about the importance of call drivers and how that’s been able to help Royal Caribbean.
[00:11:52] Chuck Baker: [00:11:52] Sure. So. So for those of you who are tunedin , our business has two primary channels, right? [00:12:00] And the, our primary channel for North America is our B2B business. And that we also have our B2C business. We work very close with our B2B partners to provide them with a lot of self service tools right and what we were not seeing is a reduction in the B2B calls coming into our contact centers. There was a lot of speculation around, you know, maybe they weren’t trained right. Maybe the B2B folks don’t understand how to utilize our systems. So this was probably one of the largest use cases that helped me to drive getting speech analytics into our business. Was building the use case around how our B2B customers are utilizing our technology and the reasons that they’re calling us.
[00:12:48] Right. So what we did was, um, did a deep dive into all of our B2B calls to understand what the primary call drivers were. And what [00:13:00] we found out is that majority of our B2B callers were calling us because they wanted to make a payment. And when we dove down into the reason behind the payment is that they would utilize or, um, or tools that are available to them to make the payments, but they weren’t able to get confirmations directly from the tool.
[00:13:22] And that resulted in us making enhancements with the tools that we’re providing them so that they could act on and make it more user-friendly or more centric to their individual needs and concerns.
[00:13:33] Nick Bandy: [00:13:33] So that whole payment opportunity led you to the realization that wow, by understanding why people are calling, we can make then effectively some changes in our operations that, so this money make us more efficient, so on and so forth.
[00:13:46] So I know one of the things early on that, um, as you look for an overall return on investment and justification to add technology was the ability to, um, look at call drivers [00:14:00] essentially as a disposition, right? Because there’s a lot of different ways to look at call drivers and we’ll really look at two of those today. One is the primary call driver.
[00:14:08] The real reason that somebody calls at a granular level. And I know Chuck, one of the things that, um, we discovered in, uh, you know, better from your perspective is when you have an agent do it, who’s got a lot of other after call work to do and has to go back and remember all the things they talked about.
[00:14:27] They have to figure out what, what was the first reason somebody called. Versus using a primary call driver or effort from speech analytics. Talk to us a little bit about how that worked out and what that was able to do for your organization.
[00:14:40] Chuck Baker: [00:14:40] Sure. So three months after launch and a hundred percent recording and, um, speech analytics, we effectively stopped all of our contact center agents from dispositioning.
[00:14:54] And for those of you guys who are in the contact center business, you understand, and can appreciate how [00:15:00] a savings of twenty to thirty seconds of every call adds up quickly. And that’s how much we were effectively able to reduce our HD just by running this out. Now getting into how, um, speech analytics do our dispositions for us was a lot of hard work around aligning what is coming from speech analytics versus what it is that was actually inside a call. That process did take us roughly about two months of constant banter in between ourselves and speech analytics while they did their fine tuning to land on the drivers, on the intent, the way the business had set it up and the way we needed to understand what each of the drivers and intents were.
[00:15:44] But that from me, um, coupled with a few other critical items that came after was one of the biggest ROIs for us. So right off the bat, within three months of launching this, we started seeing a huge return on investment.
[00:16:00] [00:16:00] Nick Bandy: [00:16:00] And from an accuracy standpoint, Chuck, from, from where you were with agents, trying to do that amongst everything else versus, um, where you are now, except.
[00:16:09] I don’t know if you want to share those numbers, but, was it a pretty big difference?
[00:16:14] Chuck Baker: [00:16:14] It was huge. Um, I’d rather not go into the granular details, but let’s say that our accuracy improved, uh, between 30 to 50%.
[00:16:25] Nick Bandy: [00:16:25] Yeah, great. Just using speech analytics. Very good. So certainly that’s one element.
[00:16:29] The idea of being able to understand right out of the gate, the first reason or the primary reason somebody’s calling, but the other opportunity within a phone call conversation, particularly, um, with, you know, in your contact center is people call for multiple reasons. And so it’s not just the first reason that’s important, but having the ability to capture all those reasons certainly has a lot of value too.
[00:16:54] Can you talk to us a little bit about that and I’ll jump ahead here and give you another side to talk to them.
[00:17:00] [00:16:59] Chuck Baker: [00:16:59] Sure. So if you think about our call intents, or even just our business as a whole, right, your, your typical customer or your travel agent will call in to say, make a payment, right. And then make a payment would be the first driver.
[00:17:15] What our agents were not very good at capturing were all of the additional intents that came in afterwards. They called to make a payment, but also call to check on dine-in or to book dine-in or they call to check on a shore excursion, uh, in addition to making a payment or they’ve called to add flights to their reservations, you know, with speech analytics, what we’re able to do is not only get to our primary car drivers on the line done accurate, but also dive more into the call to understand what are some of the additional intents that was coming in and gave us an opportunity to start looking at how we are publishing information.
[00:17:52] And I’ll digress for a quick minute. I’m just going to give you some insight into something that was huge and critical for us. [00:18:00] So we recognize that because the customers are calling about our shore excursions that we had not done a good job of putting all of the questions that customers had about shore excursions on our consumer facing website, and also on our trade facing websites.
[00:18:19] So we took the questions that were coming in as the secondary intent and built out communication and published that to help reduce the amount of calls that was coming in as secondary intents around something that we considered to be easily self serviceable.
[00:18:35] Nick Bandy: [00:18:35] So ultimately what it does is hopefully helps reduce the number of calls coming in, but also potentially reduces, uh, average talk time as well, correct?
[00:18:43] Chuck Baker: [00:18:43] That it does. Yes.
[00:18:46] Nick Bandy: [00:18:46] Terrific. And so, you know, speech analytics for the most part has a reputation of just being able to capture very large buckets of intents or reasons for calling. Talk to us a little bit about the level of granularity that, [00:19:00] um, is now achieved with Royal.
[00:19:03] Chuck Baker: [00:19:03] Without getting into a lot of specificity, I’ll talk about a couple of things that speech analytics has really helped us on over the course of the past couple of years. Um, I’ll start on one that was huge, which was when the government decided that all, um, leisure sailings to Cuba needed to end. And we really needed to quickly understand what our customer sentiment was and what their biggest concern was around that decision. You know, historically what I’ve gone around inside the contact centers and ox agents, you know, what are customers saying? What are customers saying? By using speech analytics to be able to dive into and create these subcategories, we are able to quickly go back and analyze calls and provide our senior leaders with guidance on what customers, how customers felt, which set of customers were actively looking for a refund [00:20:00] versus customers who wanted to stay on their itineraries but just understood that Cuba was no longer an option? Now let’s fast forward now to last year when we had multiple hurricanes coming into various parts of the United States. This helped us by, um, uh, by understanding, I don’t know where folks are calling from because we recorded at a hundred percent, but on what their primary concern was, instead of us just going out to make a blanket statement or making a blanket decision, we made decisions based upon feedback from our guests and from our travel partners.
[00:20:35] Nick Bandy: [00:20:35] That’s terrific. So again, I think we have, uh, keep me honest here, Chuck. I think we have about 25 buckets of intents and probably 150 to 175 more granular levels of intents that we’re able to keep track of, um, virtually on a, on a daily basis, even hourly basis, correct?
[00:20:56] Chuck Baker: [00:20:56] That is correct and that continues to [00:21:00] grow as, um, as our business expands, as we get new ships, as we introduce new itineraries, as we introduce new destinations.
[00:21:08] Nick Bandy: [00:21:08] Yeah. So I think a lot of times people think of speech analytics as certainly helpful for QA, maybe compliance depending if you’re in a highly regulated business. But I don’t know that people understand the other way areas within an organization that speech analytics can be helpful.
[00:21:26] Can you give us a quick idea of some of the other areas at a high level that, um, speech is helping in, you know, is this something that you guys look at every day? Does senior leadership have any idea of what’s going on? A little color there would be terrific.
[00:21:39] Chuck Baker: [00:21:39] Sure. So. My, um, analytics team provides that daily report to each of our brand presidents.
[00:21:49] So for those of you who ran up there we’re Royal Caribbean Group, um, is the, the name of our holding company. Under the Royal Caribbean group, we have Royal Caribbean [00:22:00] internationals, number two cruises, Azamara Cruises, Silversea plus, um, four brands that are based out of Europe. Each day, the analytics team has to provide each of our brand presidents with details on what our call drivers are, what our pain points are, what are decisions? Um, well, what are the decision trends and topics that we need to focus in on? That not only comes from the calls we’re receiving, but also from whatever data we’re also getting on social media. This has now become a staple within our business where the senior leaders look forward to this report on a daily basis.
[00:22:36] This was never present before, but now they’re able to actually view their business holistically from a 360 vantage point. Here’s what’s happening in the contact centers by region. Right. And here’s also what’s happening in the social media world.
[00:22:51] Nick Bandy: [00:22:51] Right, and you of course have the ability to filter that a number of different ways, right.
[00:22:54] By call center, by region, by whatever the case, by specific skill, anything you want, correct?
[00:22:59] Chuck Baker: [00:22:59] By a specific skill and also by where the call originated.
[00:23:04] Nick Bandy: [00:23:04] Yup. Very, very good.
[00:23:06] Nick Bandy: [00:23:06] Okay, so a lot of folks that are on the call, um, have QA responsibility. So let’s switch away for a minute, um, away from call drives, by
[00:23:15] Nick Bandy: [00:23:15] By the way, I want to encourage people.
[00:23:16] There’s a Q&A section. Feel free to, uh, throw questions our way during this. And we’ll try to address those as we move along, but let’s move the conversation to QA.
[00:23:25] Um, you know, obviously, uh, QA is a big part of what you guys do at 27 agents. Tell us a little bit about some of the complexity that you have doing QA within your organization, and then how has speech analytics helped that?
[00:23:40] Chuck Baker: [00:23:40] So I’ll talk about it from, um, you know, from where QA started before speech analytics and where we are now.. So if you guys who have QA as a responsibility to think about the transition from, from just trying to, I call it CYA. Right. Which is just making sure that [00:24:00] your agents are delivering on the policies, right?
[00:24:03] Did they do this? Did they do that to then start focusing on customer engagement and customer interaction? That is how we have transitioned our quality assurance, um, evaluations now. Where this is really about the voice of the customer. And is the agent active or playing an active role in the solution or an activity call driver.
[00:24:24] With speech analytics, what it has done is that has given us an opportunity to really focus in on the call drivers that are the most meaningful to the organization. Right. Did the cell phone calls, you know, we have the data behind them or the interactions that actually speak to the agent’s ability gives us the opportunity to filter those calls.
[00:24:44] And put that in front of our QA team members, so that they’re really evaluating interactions that we have hired our agents for. Right. How does an agent interact on a sales opportunity? How does an agent interact on a service opportunity and how do they interact with, during an [00:25:00] escalation versus going in blindly to say, you know, I’m just going to pull a call and evaluate it and hope that this is a call that really speaks to the job description.
[00:25:08] So we have moved away from transactional quality assurance and no focus on interactive quality assurance that focuses on five core competencies that our agent should have, which is really about professionalism and relationship. Being able to, um, revenue generate, being able to resolve an issue, being efficient and also selling the brand.
[00:25:31] Nick Bandy: [00:25:31] Terrific. And so, you know, we’re in a multi-channel world.
[00:25:33] So give us an example, if you can, that, you know, beyond voice where you’re able to use, uh, quality assurance or get learnings from another channel. For instance, chat and chatbots.
[00:25:47] Chuck Baker: [00:25:47] Sure. So what, what we have done with, um, our chat platform is that we have actually utilized, um, what we’re getting out of a voice channel [00:26:00] throughout and drive it into our chat channel.
[00:26:03] So we know what our customers are looking for and utilizing QA. We have used that feedback from the customers, from the phone calls and then created intents in our chat bot so that our customers are now more comfortable using a chat bot for us as calling in. So. All of the details that we’re getting through the interactions, drive some other responses on some of the intents that are now in our chat, but then again, making that channel a little bit more
[00:26:29] Self-absorbed and not having to push into the contact center to speak to an agent. So I’ll give you a couple of examples real quick. Um, you know, we, we, we had a chat bot, but, um, well when we launched it, customers were asking questions that we had not yet automated. Right. But they were going to the chat bot and ended up calling and said
[00:26:49] I was in the chat and here’s what I couldn’t find. So using that information, we pushed our engineers to make quick and critical changes there so that we could continue the chats where they’re supposed [00:27:00] to be. And the folks who really wanted to speak to an agent would be able to get through quickly.
[00:27:05] Nick Bandy: [00:27:05] Very very good. So, you know, we, we think about, um, a world where, you know, AI is clearly important, um, sort of a continued push towards automation, AI, those kinds of things. Talk to us a little bit about automated scorecards, where we are now and where, where you see that going in the future.
[00:27:24] Chuck Baker: [00:27:24] Sure. So. This year was supposed to be the year that we launched automated scorecards.
[00:27:31] And, um, thanks to COVID. We are way behind our target. However. So far we have, um, started beta testing the automated score cards and just to kind of lay the ground on where this is going, right. I don’t want anyone to walk away thinking that with automated scorecards or QA forms that the quality assurance team is going anywhere.
[00:27:54] Quality assurance team is going to stay. They’re just going to be repurposed to focus more on [00:28:00] clear coaching opportunities versus just doing the transactional quality assurance. Uh, so far we have, um, shaved off about a 30%, um, penetration in our ability to automate those scorecards. So. So far, our target is to go after the things which are more black and white.
[00:28:20] Like our policies and procedures and how those are being captured in the interactions and with all the beta tests that we’ve done thus far, Nick, we have, we are at an 85% accuracy, with the speech analytics platform, being able to capture the information which we need. What we’re working on now is the vehicle to ensure that, um, as policies and processes change internally, that we can quickly update the speech analytics platform.
[00:28:47] So that, that is captured quickly and reported accurately. I’ll add another piece to it. What this is now freeing up for us is our ability to now look at [00:29:00] our agent interaction from a 360 standpoint. So what we have started to do because of the deployment of automated evaluations is start to do what we consider to be 360 evaluations.
[00:29:12] And what that is is that we take our customer survey and then evaluate it against our QA expectations and then show the variance between how the customer evaluated the agent on the survey. For instance, how we would have evaluated that particular agent and caused us to make some tweaks, um, to the way we’re evaluating our agents, because our expectations are not always aligned with what our customer’s expectations are.
[00:29:42] Nick Bandy: [00:29:42] Great makes perfect sense. So we did a little bit of a deep dive on, on call drivers. Um, quality assurance in the few minutes we have left, let’s kind of back up a little bit, you know, the big conversation these days is all around CX customer experience and understanding that. And a lot of [00:30:00] times that’s really the point of differentiation.
[00:30:04] How does speech analytics help CX?
[00:30:09] Chuck Baker: [00:30:09] In so many different ways, Nick, I could probably sit here and talk for the next 45minutes, but the critical, um, way that it is helping our CX now is that it is has really given us a level of insight that we need around customer engagement, you know, real quickly. And I’ll, I’ll just use this as an example.
[00:30:30] We, we have, what’s known as this, um, escalation prevention strategy. Right. And this is where, because of speech analytics were able to identify calls, which has the probability of being escalated and being a publicly traded company, all of our brand presidents’ information is quickly available.
[00:30:50] Uh, they get emails and they get phone calls. So with, um, speech analytics, we are able to quickly flag calls that an interaction [00:31:00] went south and send it to a manager and or supervisor for them to reach out to that customer. Right. And trying to resolve that issue before it becomes a challenge, right?
[00:31:11] And that’s all in part. It’s to speech analytics and the ability to not only flag some sentiments, but also flag trigger words, which we have recognized in the past, which has led to escalations.
[00:31:25] Nick Bandy: [00:31:25] Terrific examples. So a couple more minutes here. Um, so what’s the feature of speech analytics.
[00:31:31] Chuck, where do you see it going from here?
[00:31:37] Chuck Baker: [00:31:37] My vision for speech analytics is that at least from my organization, it is not only being able to deploy, um, deploy the automated evaluations across the board. But I also want to be able to move speech analytics on board our ships. All right. I want to be able to start capturing phone [00:32:00] conversations, um, between our guests and our onboard team.
[00:32:04] And understand how we can quickly improve some of the low-hanging fruits, which are there. Um, and a lot of times, you know, just based upon my experience on being on board a ship, you know, they will call into the customer service team that’s on board a ship and then something gets missed.
[00:32:19] Something gets lost in translation, and I’m hoping that when we do deploy the ship board, we’ll be able to, you know, and use that to not only train some of our staff members, but understand what some of our gaps are on board the ships. I want to get to a place where the customer, where the speech analytics is following the entire customer service journey so that we can understand what happens at a booking point.
[00:32:44] What happens when they touched the web? What happens when they were on board a ship and what happens post-cruise and we can constantly improve the entire customer’s journey.
[00:32:53] Nick Bandy: [00:32:53] Awesome. That’s, that’s terrific. So, um, lots of hard work over the last six years, getting to where we are today. [00:33:00] Um, you’d have speech analytics full on for the last couple of years.
[00:33:04] Describe now, now that you have this, what would life be without speech analytics?
[00:33:08] Chuck Baker: [00:33:08] Um, I was having this conversation with, um, one of the managers last week, uh, all of the entire contact centers are working from home, right? There’s nobody that’s in our brick and mortar facility. So my question was, you know, if we had to find out how many customers are calling for a particular issue, how would we address it?
[00:33:33] Would it be Skypes? Um, to all of our agents? Or would we just pick random calls and hope that we get a good enough sample size. Life without speech analytics would probably be, um, severely compounded with headaches and, um, you know, uh, unfounded assumptions on what’s going on in our business.
[00:33:59] Uh, [00:34:00] and I think with, I know that with speech analytics, we have a firm grip on how our customers feel, what’s going on, and this is helping us to make decisions as we get ready to return to service about, you know, what’s important, what’s critical to our customers.
[00:34:16] Nick Bandy: [00:34:16] Yeah, for sure. So last question, and again, I’d encourage anybody.
[00:34:18] If you have any questions to please go ahead and put them in. Now we’re going to wrap up here in a second. So for somebody who’s in the process of considering speech analytics, um, where you were years ago, but, but sees the value. Um, what advice would you give them today?
[00:34:33] Chuck Baker: [00:34:33] Identify the use case that’s most important to your business.
[00:34:40] Right. Um, in order for me to get speech analytics on board, I had to do just that and Nick can attest to this. I was only able to afford, um, five thousand hours of speech analytics for the first year, but that gave me enough, um, bandwidth to be able to chase the use case [00:35:00] and tie an ROI to it. So once the use case that you have identified as critical enough for the business and you can show what speech analytics can do with whatever you’re able to get with your pilots.
[00:35:13] I think that would be huge and it will send the right message and everything is also about timing. Right? Just got to ensure that whenever you’re positioning this, whenever you’re bringing it to your senior leaders, that the timing is right.
[00:35:28] Nick Bandy: [00:35:28] Yep. Yeah, very good. So Chuck. We did get one question from Chris.
[00:35:31] How do you measure the accuracy of the automated scorecards? Do you trust the results?
[00:35:35] Chuck Baker: [00:35:35] Sure. So what I do know, right, enough for us to, um, to move forward is that I do on audit, right? So we, we bucket our calls into 200 calls each time we are running our tests for each of the various attributes, I have my quality assurance team go in.
[00:35:57] And evaluate 200 calls and then run [00:36:00] them through the automated scorecard. And then we compare our results to what the automated score card delivers and see where the variance is. Now, let me start this by saying, when we first started doing the automated, um, scorecard, accuracy was 20%, right. And it took a lot of tweaking and enhancement on the SpeechIQ side, and a lot of, um, changes to their word recognition and them
[00:36:26] boosting their fanatic engines within their platform to get to the 85% that we have now. My target is to get to 92%. So there’s still a lot of work in progress but I acknowledge that working from home has its issues.
[00:36:44] Nick Bandy: [00:36:44] Very, very true. So it doesn’t appear. I’ll check one more time for any more questions.
[00:36:49] It doesn’t look like it’s. So, um, quick FYI for everybody, there will be a speech IQ demo. Sherry, if you can share with us the time. Uh, the date of that. [00:37:00] That we’ll drill down a lot more into some of the things that Chuck talked about.
[00:37:03] Sheri Greenhaus: [00:37:03] Yeah. The demo is this Friday, uh, at 8:30 so it’s a half hour demo, but, um, Chuck, I have, I have a couple of questions for you.
[00:37:12] So I’m assuming there are a lot of QA analysts watching today.
[00:37:18] How has their life changed?
[00:37:24] Chuck Baker: [00:37:24] So, I talk about how my QA team’s life has changed and what I’ve done to prepare them for their new role. So my QA analysts are moving from, um, being an analyst to now being a coach. Right. So what I’ve done, what they’re doing now, bear me for a second. I’m sorry.
[00:37:48] Sheri Greenhaus: [00:37:48] You’ve been speaking a lot.
[00:37:55] Chuck Baker: [00:37:55] Apologize. So my QA team [00:38:00] members are changing from just being focused on evaluations to being focused on the business holistically. So, not only are they now analyzing the data that’s coming back from speech analytics and providing that type of results to the senior leaders, but they’re also acting more as coaches to help the supervisors to improve their agent’s performance.
[00:38:19] Um, the strategy from my quality assurance team and the businesses is that my QA team members don’t coach agents, they coach supervisors on how to provide feedback. Because I want that responsibility to continue living with the supervisors because they’re responsible for their team. So now, the QA analyst has more time to spend with the supervisors going over data that’s coming in from your teams and start to identify how and what needs to be coached.
[00:38:47] They also use the data to also drive training. So as they see areas of opportunity that come up, they engage with senior developers, um, for the training department to start working on training snap bites. [00:39:00] that will impact certain areas of improvement for the agents and get those out. So they went from just providing data on agent’s performance to now providing data, not only on the agent’s performance, but also on business performance and clear training opportunities.
[00:39:18] Sheri Greenhaus: [00:39:18] So from, from that, how does that then change how the supervisors use the analyst when they come with recommendations for them?
[00:39:27] Chuck Baker: [00:39:27] That was, um, that was a transition. So first, you know, it was, it was QA versus the business at one point. So we, we had to build that bridge, so that the supervisors started to understand
[00:39:42] How the QA role is changing and that came from collaborations. And that came from there was, um, meetings and sessions with them to talk about our transition. Um, three years ago, we started making that overall transition. The QA team, you know, met and did a presentation to the business to [00:40:00] say, here’s where we are now.
[00:40:01] And here’s where we’re headed. And these are the steps that we’re going to take to ensure that we’re there to support you. Clearly some of the other things that we have done is that our QAs are now considered to be, um, parts of the transition team for new agents coming into our business. Right? What they do is that when agents come out on new hire training, they act as, um, coaches and they help with, they’re now responsible for a transition.
[00:40:30] And when they transition a new team over to an agent, or over to a supervisor, they make sure that all of the agents are performing with an 80% of their established KPIs. So now the supervisors are more engaged in what the QA team members have to offer.
[00:40:47] Nick Bandy: [00:40:47] Hey, Sharon, we got a couple of questions then from the folks watching.
[00:40:52] Yep. So, um, one question here Chuck is, uh, did you need to change your evaluation form to improve your predictive score? [00:41:00] Some I’m presuming that means your auto evaluation forms. How did you get to better accuracy?
[00:41:06] Chuck Baker: [00:41:06] Actually, we have not made changes yet to our, um, QA score. SpeechIQ is the one that I’ve been adopting their platform to meet our expectations.
[00:41:19] No, I, I say that that is the way it is now but as we dig more into our evaluations, you know, I think within the course of the next couple of months, we are going to be presenting to the business that there needs to be some shifts with the way we are evaluating our business. Uh, and then we’ll probably, we will definitely do some changes to the quality assurance form.
[00:41:39] Nick Bandy: [00:41:39] Great. And I think that spurred a few other questions here. Did you have to adjust your analyst team size and how long did it take you to train them to transition from analyst to coach?
[00:41:47] Chuck Baker: [00:41:47] That training is still ongoing I should add, um, we, we. When that process initially started, [00:42:00] it was about doing a gap analysis to understand, um, how the current head, the current folks that were in the QA organization could then match up to the new expectations.
[00:42:11] So we wrote the job descriptions and did a gap analysis and then identified what training needed to occur. We did have some fallout, I think we lost, um, two persons who decided to pursue other, um, opportunities within the company. Right. But for the most part, they started off with just 27 after the two persons walked away.
[00:42:35] They went through a series of training, focused on, um, understanding analytics because they went from just doing a QA form to actually utilizing data and presenting it. So they had to go through project management training and they also went through relationship training and professional leadership training.
[00:42:52] So that training lasted on the average of about six months. It was not everybody all at the same time. [00:43:00] I think I missed one part of that question.
[00:43:05] Nick Bandy: [00:43:05] I think you got it. Did you have to adjust the team size and how, how long did it take you to get them to transition?
[00:43:12] Chuck Baker: [00:43:12] Right. But I did not trust the team size because what was critical for me was not to lose headcount. Right. And I didn’t want any of my QA team members to feel that automation.
[00:43:22] Was going to cause them to lose their jobs. Right. So what we did was started to identify how else we could utilize the QA team members across the organization. And there’s so much data there for them to play with, so much things for them to do with that. I wanted them to have the ability to actually drive how speech analytics impacted our business.
[00:43:44] Nick Bandy: [00:43:44] So there are a number of questions coming here that, so I don’t want anybody to think that I’m missing you, but there are a number of questions that are sort of similar to that. Well, um, you know, people are saying that our agents are good at scoring and coaching, but they’re not necessarily highly skilled in [00:44:00] analytics.
[00:44:00] So I think you addressed that. It’s a journey trying to kind of pivot and not everybody’s going to make that journey, but ultimately the organization has to move along and you may have to find, like you said, a couple of those folks opted for other areas within the organization and analytics just wasn’t where they wanted to be. Right.
[00:44:19] Chuck Baker: [00:44:19] Correct. I mean, and there were some folks who really embraced it and capitalized on all the training that we provided. I think from me, and I think this is something that guys should, you know, if you’re thinking of going down this road really try to provide your folks with that opportunity. The gap analysis helped us a lot, because it’s really quickly identified where we needed to focus our training on, um, once you
[00:44:49] present your journey to these folks and get them involved in the initial phase, when we were doing the, um, use cases, everybody on the QA team was [00:45:00] involved. Right. So they knew that this was something that was common. This was not a surprise. So they became part of the journey. Yeah.
[00:45:08] Nick Bandy: [00:45:08] Yeah Yep.
[00:45:09] So a question back to call drivers, somebody asked was the primary, uh, call driver concept, the key item that drove you to seeking a speech analytics tool, or was that a secondary item to other, uh, speech analytics capabilities?
[00:45:24] Chuck Baker: [00:45:24] Well, the call driver was secondary, right? Uh, like I said, I was trying to look for more efficient ways for my QA team members to do more evaluations. I was trying to get to more data quickly so that we could, you know, shape the business and help improve performance. As you think about the amount of calls that we take on an annual basis.
[00:45:47] You know, we were only scraping the surface with all the calls that we had access to. And, you know, whenever somebody asks about something, you know, it meant that we have to go try to find a call [00:46:00] that we hope could help. You know, sometimes our senior leaders would reach out to us and ask us for calls on particular areas.
[00:46:08] Then we would have to sit there as a QA team and listen to calls for two days before we could find it. So we were just trying to find ways to be more efficient. And as we started diving into speech analytics on the hundred percent recording, that’s when I was like, you know what, maybe it could do this.
[00:46:22] Maybe it could do that or can it do this? Nick, can it do that. Nick. If I asked you to do this, would they be able to get that done? And it was, it’s just like when you get data, you start to, you start to do your pivot tables and you started to see trends. You’re like, oh, I didn’t know that was there. Let me see if you can do this so the more you dove into it, the more I dove into it.
[00:46:42] And the more the team dove into it, as more, we recognize that there is so much more that we could get out of it. So we started seeing how it could work.
[00:46:51] Nick Bandy: [00:46:51] Yup. Makes great sense. And finally, somebody asked, do you see AI as the future of contact center QA?
[00:47:01] [00:47:00] Chuck Baker: [00:47:01] I do, just for agent evaluations, I do see AI as playing a significant part in evaluations. So I’m saying that, um, because I believe that that is the direction we need to go in, but I don’t think that AI can stand alone. To be able to provide a level of, um, coaching performance improvement that’s needed for agents to be successful, I think that’s one of the vehicles.
[00:47:29] Nick Bandy: [00:47:29] Yep. I would, uh, I would agree with you. These are terrific questions. So, uh, I appreciate everybody who, uh, uh, sent us a question. I think it made for some really interesting conversation. Um, I would encourage everybody, please stop by the speech IQ booth, where you can learn a little bit more and chat with somebody who can drill
[00:47:46] You know, at a deeper level at some of the concepts that we talked about today, Sherry, thank you for putting on this terrific conference. And Chuck, thank you for all of your time today. I think it provided for some really, really good [00:48:00] information and conversation and, um, your cruise, uh, uh, ship.
[00:48:06] Views at the beginning were very inspirational on an icky day in Ohio. So thank you for that. And I hope everybody got a little bit of information and ideas out of the chat today.
[00:48:16] Chuck Baker: [00:48:16] Thank you.
[00:48:18] Sheri Greenhaus: [00:48:18] Thank you both. Yeah, Chuck, it was great to see that because we were, I was getting tornado warnings on my phone. So that looked really good.
[00:48:27] Um, just as a reminder is you have not yet registered for the demo on Friday. There’s still plenty of time to do that. And that’s at 3:30. And, um, SpeechIQ is also going to be, uh, one of the sponsors of our magic show, which is tomorrow at two o’clock.
[00:48:45] So if you have not registered for the magic show, go ahead and do that.
[00:48:48] And of course go over to the booth.
[00:48:51] So thank you, Nick. Thank you, Chuck. Thank you, everybody that attended. We do have a session
[00:48:57] At two o’clock today and [00:49:00] I will get the title of that is, um, the twenties rethinking your quality program and the time of dramatic change. And there is quite a bit of dramatic change.
[00:49:10] So we hope to see you there at two. So thank you everyone. At this point, we will be closing this session. It is recorded. If there’s others you would like to have see the session or listen to this session, it will be available in about 24 hours. So thank you all. And we’ll see you at two.
[00:49:29] Nick Bandy: [00:49:29] Thanks everybody.
[00:49:29] Take care.