Today, the path to winning and retaining customers requires a unique approach to customer service. But without the right data and tools, contact centers are unable to create memorable customer experiences.
In this webinar, our panel featuring guest speaker Michael DeSalles, Principal Analyst at Frost & Sullivan, discusses:
- CX trends and predictions that will change how contact centers deliver differentiated customer service
- How to implement the tools and capabilities necessary to create positive omnichannel experiences
- Ways to easily leverage data, practical AI, and workflow automation to achieve meaningful digital transformation.
Nick Bandy: [00:00:00] Hi everyone. And thanks for joining us today. My name is Nick Bandy, CMO at LiveVox. And joining me are Michael DeSalles, Principal Analyst at Frost and Sullivan and Boris Grinshpun, General Manager of CRM and Messaging at LiveVox.
Gentleman, thanks for joining today.
Our subject is ”Delivering Unmatched CX Through Seamless Customer Journeys and Practical AI”. It’s a certainly very relevant topic. Um, so let’s dive right in. Michael, I know Frost & Sullivan has done a lot of research in these areas. So please share with us some of the findings that you have.
Michael DeSalles: [00:00:34] Well, great.
Thanks Nick so much. I’m delighted to be a part of this webinar and I want to thank you and Boris, and of course, the folks at LiveVox for the invitation to share these research findings, as well as commentary from interviews we have with contact center leaders. So as we bring up the first graphic, uh, I wanted to highlight what we’re seeing as the main challenges and issues that are having a huge impact on operationalizing a contact center.
And actually I prefer the term interaction center, or even better relationship hub, that was coined by Bruce Temkin, a customer experience transformer. I wish I had thought of that myself, but I wanted to point out just a few here. So apart from the already difficult task of recruiting, testing, and hiring new representatives, it’s the contact center’s struggle to train and more importantly, retain agents that they already have. My comment is that agent attrition, which can range from 50 to 80% of the contact center is what I consider the scourge to the industry and one of the biggest problems. Agents even today have too many screens and systems to access, to address even the simplest inquiries.
It might be billing, account balances and changes, maybe even package delivery status. Well, it slows everybody down and adds to my frustration as a customer. Although I think it’s gotten better over the years, we all continue to have to repeat our basic information, either online or with a live agent and then tolerate long hold times.
And then finally, it’s really hard to keep track of all the moving parts. In running a call center. If you look at digitizing it, innovating, and staying current with new tools, new platforms and systems that are in front of, alongside, and behind the agent. Next slide, please. So here’s some of our data. And from an empirical standpoint, the next graphic shows the results of our survey of over 3000 customer experience executives that we conducted earlier this year and asked this question.
Which key digital transformation objectives have been accelerated as a result of COVID-19? Right. And as you can see from the top three in the ranking percentiles are actually pretty close here. They include adapting to new work modes, which I think we’re seeing. And that could mean remote working, staggered or modified shifts, uh, this new hybrid work environment moving towards greater digital customer engagement on the self-serve front end. And I think we’ve all experienced that.
And of course, enhancing e-commerce capability. I think it’s important to note that very close behind all of these rankings are one — improving client engagement and customer experience along with increasing capacity to respond to higher customer demand, which we might see with our economic recovery just on the front of things.
Next slide, please.
So we might ask, how does all of this play in the context of disruptive new technologies? New business models in multiple verticals, verticals, and geographies. Well, this next graphic shows how we at Frost & Sullivan conceptualize this radical change happening in our world. I think we can reasonably conclude that everything from new business models to disruptive technologies are having impacts across all industries and across all geographies.
And if we can imagine each ring in this circle, kind of turning and brushing against one another above it and below it sort of like tectonic plates, this kind of world spinning with changing dynamics, the process of technology and all the data that comes with it are being produced in our world is extremely disruptive.
Partly what I find so fascinating about this is we’re living in a world of immersive technologies and of course, then customer experiences. And that means to me, that within the next three years or so, the exponential growth of phenomena such as digitalization and machine learning will fundamentally change how businesses create value.
How we satisfy customers and outperform competitors, but two things jump out at me here. First, companies have to take actions that really position them for the next level of success. Understanding which digital technologies are required today and which ones will be important tomorrow is critical to make effective plans, to prioritize budgets.
To sequence investments and schedule implementations in order to maximize revenues companies that don’t embrace digital will probably experience what we call a Kodak moment. And then second, I’d like to suggest that the goal here is to develop an agile and digitally transformed enterprise that’s capable of acting and reacting to all of the perpetually changing world events, this massive data lake that we all deal with and consumers with a whole different set of values, beliefs, attitudes, and maybe even lifestyles.
Next slide, please.
Nick Bandy: [00:06:14] So, so let me jump in here. That’s a terrific slide. There’s a lot of moving parts in there. Um, so Boris quick question. When you, when you see the slide you think about, um, you know, obviously a lot of new channels, digitization in data, how do you think about how all that comes together from a, from a product perspective, you know, what comes to mind for you in that.
Boris Grinshpun: [00:06:38] Yeah, I think that’s a, that’s an excellent question, Nick. I think what we’ve seen sort of over the course of the last number of years is sort of this explosion of ways that consumers are looking to communicate. And what we’ve seen over that course of time is we’ve seen many organizations sort of add on these additional communication methods to get to a place where they’re communicating with the consumer and essentially servicing the consumer in that specific channel.
And that may be sort of like the first foray, but I think what’s been very challenging for a lot of uh, a lot of clients, a lot of enterprises today is to be able to tie all of that communication in a cohesive manner and to be able to understand the full communication journey that occurs between the enterprise and the customers.
What we’re seeing, what we’ve seen over the course of that time is that enterprises have run to meet the demand, but it hasn’t been really thought through strategically and therefore we’ve seen a huge proliferation of silos. Uh, you know, just to service a particular channel or just to service a particular business unit.
Michael DeSalles: [00:07:55] Great. Thanks. Let me, uh, comment on this next graphic. So as we look at this whole host of industries that are affected, here’s some examples of how customer expectations might be changing for certain industries. For retail, for example, it’s all about creating this highly personalized customer experience across all those channels that Nick mentioned and devices.
It can be in clothing, electronics, even buying tires, which I experienced recently. You’d be quite surprised by how personal it is. If you look at banking and finance and consider maybe on the business, as well as the consumer side, mortgage lending, credit cards, student loans, wealth management, it’s driving a seamless experience across each one of those very discrete activities and then healthcare, which is so topical today.
We see the providers, payers, and patients expect integrated, intelligent patient care systems. They want telehealth options and healthcare data management to support what I call a fluid data movement, very flexible access interoperability, and a Swift scale up of innovative applications. They’re there to support patient interactions.
Next slide, please. So here we are with Frost and Sullivan’s definition of omnichannel, which frankly, we coined back in 2013, eight years ago. And it hasn’t altered very much, hasn’t moved. You know, the first half is pretty well-known, you know, omni-channel is ensuring a consistent and seamless high quality customer experience.
But the last sentence to me really puts the emphasis on the why, why is it so important? And that is that it ensures that data and context from initial contact carries over to subsequent channels, thereby reducing customer effort, improving the customer interaction and then enabling the business to tailor the customer journey.
I’ll say just one last thing on this. When it comes to discussions and thought leadership, you know, rumors about the term omnichannel going away and the dying, maybe in favor of some other definition are greatly exaggerated. What I want to reinforce here is the importance of delivering a consistent omnichannel experience, which enables contact centers, those relationship hubs I referred to earlier, to improve every single area of their operational strategy.
Nick Bandy: [00:10:40] So how has that definition evolved? That’s interesting.
Michael DeSalles: [00:10:43] Oh, I have something on that. Thank you. So, so Nick, while we’ve not changed the definition we have added to it to include artificial intelligence, AI-enhanced tools, knowledge management, and the impact on interaction outcomes in generating the important insights that actually improve efficiencies, agent engagement, which we all know is very important, in order to close the loop on comprehensive customer satisfaction.
So. Forgive me for reading, but here’s what it sounds like. Omni-channel enhances the bridge between the customer and the agent experience, incorporating AI-enhanced tools and knowledge management to further improve the outcomes for both. And the resulting data generates deep insights, to assist in continuously improving operational efficiencies, agent engagement, and ultimately customer satisfaction.
Next slide, please. So what is an omnichannel customer strategy really aimed to do? Well, one is to improve the agent experience, hence the customer experience. It does help you get to that higher level of personalization because at the end of the day, in the customer’s mind, they’re asking how well do you know me?
And it moves you closer to that sometimes aspirational and hard-to-achieve goal of consistent, seamless customer experience. I mentioned AI, but it further enables AI and automation by driving simplification and provides a competitive advantage. Given the pace of change in the world of customer contact.
Next slide please.
Nick Bandy: [00:12:30] So yeah, but that last point, um, where we focus on AI, you know, clearly that has an impact. And you’ve said that you’ve, you’ve, reorchestrated your definition of omnichannel based on it. So let’s spend some time now. I think it’s a topic everybody’s kind of really interested in, right?
Artificial intelligence. Um, what to do with it in the contact center. Um, how to get started. So let’s shift the conversation there and I, and I see on our next slide, we’re going to define a couple of things within AI to make sure we’re all talking from the same page. So, so Michael, please share with us that, and then we’ll, we’ll transition the conversation there.
Michael DeSalles: [00:13:09] Thanks so much for that Nick, but I wanted to take a minute just to level set, ensure our definition of artificial intelligence, you know, as we use the term and then make maybe a little bit of a distinction between the use of bots and intelligent virtual agents. So very simply artificial intelligence for us, you know, seeks to emulate human cognitive capabilities and importantly assist in decision-making with high accuracy and speed using data-driven intelligence and self-learning abilities.
Now, bots, again, I kind of take responsibilities and analysts in the industry for the confusion between bots and virtual assistants. But let me say that bots are computer programs and they’re really built to engage with an individual and with the artificial intelligence, then emulate humans through either web chat or speech interfaces. You know, they can be basic apps that can answer simple inquiries, or they can seek to entertain, to fully conversational bots — the advanced ones with intelligence that’s embedded into the app and the ability to be integrated into backend systems. Then finally, we’ve got intelligent virtual agents. So these are speech or web enabled applications, and they’re backed by sophisticated AI technology, emulating humans in a customer service environment.
For example, like agents in contact centers and they can be speech or text driven or a mixture of both. I think we’ve all experienced that. And in our opinion, intelligent virtual agents and bots differ in terms of their scope, their complexity and capabilities, with virtual agents being more sophisticated in my opinion.
Next slide, please. So you might ask what are some of the top reasons that companies are investing in artificial intelligence? Our survey data shows in very close percentage ratios, one — to improve the quality of products and services. Two — to keep up with and get ahead of competitors. And here’s a very important reason, to automate customer contact functions by reducing those easy transactional call volumes to agents, improving efficiencies, and finally lowering costs.
Next slide, please. So this last look at our survey data shows that there are a number of benefits to be derived from using AI. And there are a whole host of reasons, but the most important ones for contact centers are guided assistance for agents and back office workers and at the heart of it, enhanced customer experience.
So listed here in this graphic are just some of the many benefits that can be derived when automating customer interactions using bots and virtual agents. And that would include the data, the analytics and the AI behind it, but again, to go to improve the customer experience. So here’s some of the benefits 24/7, always-on, quick automated responses.
This frees up agent time for more complex tasks. It creates those personalized experiences I mentioned earlier and then enhances self service and refines the customer journey. But let me also mention just a few of the benefits for agents. If I’m an agent, it gives me the power to help my customer better. It makes me feel more secure and confident that I know who they are when they contact me.
Agents need assistance with research so we can be a research assistant. Very important, especially in financial services and healthcare, it keeps me in compliance. It helps me be reminded of certain tasks that have to be done. And then finally, it’s an agent that assists me with call wrap-up so I can get onto my next customer.
Well here, I like to recap to close out. Let me just talk a little bit about the how and why when we consider the advantages of virtual agents. We’re using natural language processing algorithms to understand verbal nuances and I think that the technology does a great job of that today because it engages in human-like manner, making even more difficult customer transactions, which can be language based, uh, they can be a regionally based, more personalized and more conversational.
And then finally, there’s built-in machine learning, that virtual agent actually gets smarter with every interaction and it’s very easy to iterate on without engaging IT. Virtual agents, and this is for customer satisfaction, offers very fast routing. Well, Nick, that concludes my portion of the presentation. Let me turn it back to you.
Nick Bandy: [00:18:19] Really good stuff, Michael. Thank you. So Boris let’s, um, let’s turn this over to more of a perspective of product management and what you’re seeing day to day within end use customers and, you know, focusing again on this whole notion of practical AI.
Boris Grinshpun: [00:18:35] Absolutely.
Nick Bandy: [00:18:36] But beforehand, why don’t you give us a quick 30 second overview of LiveVox?
Boris Grinshpun: [00:18:40] Yeah, absolutely. Um, for folks who may not be familiar, uh, with the organization, LiveVox is a leading Contact-Center-as-a-Service, uh, software provider. Uh, today we help our clients communicate with consumers and prospects over 14 billion interactions.
Uh, we of course have been a number of leading contact centers in the enterprise space, utilizing the platform. And, uh, the platform is really a full suite approach to offering all of the right tools and components for your contact center to be modern and, and optimize every consumer interaction across channels, CRM conversations, as well as a full workforce optimization suite.
So that’s a little bit about us, Nick.
All right. So let’s talk a little bit about the lens. So we hear a lot from customers today. We’re really looking to dive into the notion of utilizing AI and really oftentimes may not know exactly where to start. Um, more specifically, there’s sort of a bevy of things they’re looking to automate.
Um, there are a bevy of things that require additional assistance. And so we always get the common question of, well, where do I get started? What should be the first things that I look at? And we’d like to boil it down for a lot of clients in terms of, well first looking at the overall customer experience.
In other words, what are your customers looking to do? What are they looking to achieve? What does a successful transaction look like for them? That’s sort of number one is identifying the, I’ll call it the customer or the end consumer use cases that you’re looking to automate. The second sort of lens is the technical feasibility of executing such events.
Now, um, I always like to make this practical. What do I mean by that? Well, if you ask the customer today, um, Hey, what would you like to do? The customer would raise their hand and say, I’d like everything to be automated. I want everything, in an easy to manage manner, but the reality of it is that some things are not feasible, uh, technically even for your organization.
Uh, I can give you an example of that. Everybody would love to say, “Hey look, can I please, uh, you know, refinance my house since interest rates are, uh, are very low, but the actual technical feasibility of executing that is sort of very, very low. On the reverse side, when you look at the technical feasibility of a common interaction, like resetting your password, you of course can execute that there are certain events and APIs that you can call.
So the probability and the success factors are sort of very, very high. The third capability is the propensity to complete the interaction. So commonly, when we have conversations with folks, uh, on the enterprise side, they would like the customer to do X, but the customer doesn’t actually want to do that, whether that’s a channel constraint or whether that’s sort of, the feasibility of doing something, it just doesn’t, it doesn’t sync.
So you may have a customer that maybe wants to obtain a new loan, but the possibility or the probability that they would go through, let’s say a 20 to 30 minute journey, uh, with a virtual agent and have a back and forth conversation, is sort of not probable. And so this is where we talk about that third leg of the stool, which is what is the probability of completion of sort of, of sort of these events.
So when you put these three events together, they sort of should form a lens for you of, is this an event or is this a component that I can use a virtual agent in order to automate?
Nick Bandy: [00:22:41] And so basically that element of self-service.
Boris Grinshpun: [00:22:42] Yeah, exactly, exactly. And, um, so before we sort of, uh, even go even one level deeper in it, um, let’s not forget why we’re doing all of the, all of these things. I know Michael talked a lot about this, but I would just want to sort of hammer home the benefit of this and it all starts with self-service. We have taught, or, you know, the likes of large tech companies have taught the consumer today that it all starts with self-service.
It all starts with you Googling or searching something on the web. With the proliferation of voice devices now in your home, a lot of folks have been taught to just say, Hey, why don’t you just tell me what it is that you’d like, uh, help with. So it all starts with really sort of that notion of self-service and for high volume contact centers today, um, this helps them essentially provide that certain self-service capability, especially in events when there’s high call volumes.
So if you’re sitting here today and thinking about what, where is this going to help me? Well, it’s going to help your contact center not only deliver on the self service promise, but it’s going to help you deliver on that self-service promise, especially when you have high call volumes and you know, there’s this sort of threat that the customer is going to be in hold.
So that’s one of the ways that this helps your contact center today, and of course reduces your abandonment rates and it gives customers what they want, which is that 24/7 self service capability.
But I think the last one is very, very important. Michael mentioned this a little bit earlier on in the conversation. You mentioned a very interesting metric regarding, um, regarding the attrition that a lot of contact centers exhibit today. And this is commonly overlooked from the benefit of a virtual agent assist model.
And this is the capability in order to lighten the load of your contact center agents from doing very simple, repetitive tasks. Those tasks have commonly caused high levels of attrition within our workforce. And it really sort of drove down or have really driven down the pleasure, increasing the displeasure of the job.
So these are sort of the ways that you can think about the immediate benefits that this will provide, or these solutions provide to your contact center today.
So as we look at things, I think it’s always very practical to give folks a little bit about, um, a view in terms of what are some of your easy wins. And so we kind of broke these down. I’ll kind of call it the good list versus the not so good list. The good list is in the blue, uh, for folks, uh, you know, uh, for, from a color coordination standpoint.
So what are good things? Okay. Executing on a payment transaction, very structured, a very high probability that a customer’s going to finish, um, very short in its nature and very confined. It was a good sort of a, if you will, reason to deploy self service. Um, checking out an item and its status or, uh, activating a certain transaction or a card or rescheduling an appointment.
All of those things are really great things to automate now. One of the things is not so great at automating — whenever there are nuanced conversations to be had. Whenever there’s a lot of choices, whenever there’s a visual element that needs to be added to the interaction, these conversations aren’t ideal.
And so we get a lot of questions from customers that say, well, do you know customers call in quite often to make a complaint regarding a certain product. Let’s automate that. I would stay away from those types of automation requests, because you’re not likely to have success. You’re not likely to achieve your goal of delivering great customer experience.
And there’s likely to be a lot of ambiguity, and there’s going to be a lot of customer frustrations. Customers want to be heard. They want to receive that empathy and no matter how good your virtual, uh, virtual agent is. Uh, we know one thing: they’re not very empathetic. So, um, I would focus on the blue, keep the small, keep this in a box and these are probably your best and easiest wins.
Next slide please. Okay.
Boris Grinshpun: [00:27:17] So I think, you know, throughout this conversation, um, we talked a lot about the interactions. We talked a lot about what can make them successful. And I don’t think any, you know, this is really all that debatable, but the question becomes from an industry standpoint, why is this so hard to execute?
And the number one reason that we see that the execution of this is very hard or people have sort of this. Um, that trough of disillusionment, if you will. Um, the problem is that largely people will look to implement a virtual agent capability, either one in a specific channel or only over an incomplete data set.
So let me talk to you a little bit about that. So first let’s talk about the incomplete dataset. Um, we know we have multiple conversations today with our brands or the brands that we love. I think of these little bits and pieces of data located in many disparate systems. And so when you start to have a conversation with a virtual agent, the moment that that conversation can’t be personalized, the moment that they get something wrong, the entire trust factor breaks down completely.
So when we start to have conversation with clients today, one of the first things that we talked to them about is, well, where’s the data storage going to be. Where’s that data storage that we’re going to tap. Yeah. Um, more often than not, we know that our clients are having to aggregate and centralize this data in order to have a successful deployment of the virtual agent or bot.
And the lack thereof leads to really an unsuccessful deployment, uh, of, of these tools. The second thing is that people tend to think about virtual agents or bots channel-specific. And I think sometimes that may be, that may be right, but at times it may be wrong. Here’s where it may be wrong. We know that a lot of conversations today start through the utilization of digital channels.
A lot of people are, you know, have a mobile device they’re looking to get self-service. The conversation may start via web chat, it may then sort of transition to a voice channel if it needs to be escalated. Unless there was sort of some contextual awareness between that switch happening and that agent knowing what, uh, what has happened in the past.
This is when things sort of tend to fall apart. And this also starts to balloon on an organization, the number of virtual agents and bots that they’d have to deploy. Deploying a different one for every channel with a different data set can quickly become a mess. So I guess what is, what is my point here?
My point being here is that the more you can centralize or have a central repository for data, the more successful you’re going to be in deploying these tools.
Nick Bandy: [00:30:15] Hey, Boris, we’ve got an interesting question that just came in. Um, and that’s obviously related and they asked as NLP gets better. Um, does that ultimately help improve the overall accuracy and capabilities of, uh, of your, your IVA?
Boris Grinshpun: [00:30:33] Yeah. I mean, absolutely. Um, it’s actually a really good question. As NLP does get better, I think you have a higher probability of having an understanding of an intent that our customer has whenever they’re contacting your organization. So in that particular sense, uh, you’re less likely to put somebody down a particular incorrect path.
So certainly, uh, certainly that’s, that’s the case. However today, when we look at the challenge of today, we actually look at the challenge much more from a backend process standpoint. Today, NLP is actually pretty darn good. Um, there are many different ways that you can get analytics and data, the training tools that are out there are amazing.
And I think we’re going to talk a little bit about that in a moment here. But I think the largest challenge for organizations getting here today is, um, how can I make this bot not just have a routing conversation with my customer, but actually tap into a consistent data set so that the conversation is personal and that I can execute certain functions that my backend systems are ready for all of that.
Nick Bandy: [00:31:43] I’m hearing a consistent theme here, having that ability to have a unified data set and customer data is, is really going to be key to the success of this, um, and the ability to scale it. Okay, great. Let’s keep rolling.
Boris Grinshpun: [00:31:54] Yeah, absolutely. So, um, so why is this important? This is actually, um, pretty important. I think Nick you kind of took me down this path. In terms of, well, what happens when you actually get to the point of deploying a virtual agent or, or a bot? Well, I think your journey doesn’t exactly stop there. And I think a lot of people kind of forget about this.
Your initial deployment may go, well, it may go very so-so, but those should be your expectations. Because one of the things that you’re going to do very quickly afterwards, A, to just check your work, but B, would be to make sure that you’re going to have success is you’re going to actually invest a lot of time monitoring these agents and the conversations they’re having with consumers today.
Um, you’re going to want to make sure that people aren’t frustrated, you’re going to want to make sure that, um, the way that you sort of verbalize something or the way that you’ve structured your intents is the way that customers, are going, if you will, through those particular, particular flows.
This is an example, um, out of LiveVox platform in terms of, uh, how we provide the capability for folks to monitor, uh, virtual agents. Of course, lots of other companies in the space have other capabilities, but having the ability to in real time, tap into a conversation will prove immensely valuable to your organization in improving this.
And I think this is what you want to do, you know, whenever you’ve set out any technology project, is to test, how is it actually performing in the, in the field?
Move on to the next slide. What you’re also going to want to do is what you’re going to see, and you are going to want to improve on your intent.
So Nick you brought up an excellent point a little bit around NLP and, um, what’s interesting is of course, we can understand things around the edges. Yes. Yeah. Okay. That’s fine. All of those things are already largely pre-baked in a lot of models, but what’s interesting when we talk to a lot of organizations, there are assumptions regarding the way that customers communicate problems or the way that they call certain things and certain businesses may not be what you currently have is your own taxonomy.
And, um, One of the best ways to check that taxonomy is to constantly be iterating regarding your flows and being able to visualize them to say, Hey, look, people are kind of going down this path and we have some really good answers for them, which are, which is excellent.
Uh, or folks, are kind of going down this path and maybe it’s not an ideal path. Um, maybe we’re intending for them to do something else. And so having the visual map where customer intents and conversations are balanced against each other and utilizing tools like, um, like historical data sets or in this example, the speech analytics tool gives you that sort of great insight.
So that way you can take those back to management and say, “Hey, look, we need to do something different. We thought people would, would self-service or will not self-service and we’re getting a lot of either demand or we’re getting a lot of friction in this particular point”. So this is a little bit of a best practice around how you can do that.
Nick Bandy: [00:35:09] So as speech analytics becomes more and more commonplace in contact centers and as people, you know, try to find a return on investment. In doing so, this is another way where speech analytics pays off is what you’re saying. This helps inform the bot, and helps you get better and better at it, right?
Boris Grinshpun: [00:35:28] Absolutely. And this has been a, really a, sort of a great tool because whenever we have a lot of conversations initially with clients, they kind of I’ll call it. They realize that the bot is this black box, right. Something goes in, the bot kind of turns through that, gets very, not very frequently tested and out comes some results.
I think that’s, um, maybe. Maybe the IVR way of looking at things, if you will. Um, we see a lot of forward thinking organizations having moved a little bit past that and saying, okay, it’s not enough that I just did this. I actually need to manage this on a fairly frequent basis and make sure that I’m making tweaks and iterations.
And I have all the right tools to make tweaks or iterations. Otherwise quickly, the value of this is going to decline over a period of time.
Moving forward. And this kind of breaks things down a little bit further. Um, so that way, uh, you have the right data that you understand all of your tests. You also understand the frequency, um, all, all of these intents that are happening. Um, I’ll give you kind of a little bit of a data point when we’ve deployed this to a number of clients.
Uh, we usually will start sort of with a set number of flows, and what’s pretty interesting about this is the initial assumptions around what customers will want to do versus what they actually end up doing. And I think this is very interesting because this is sort of the data set that you should be managing within your organization.
It’s not enough just to make assumptions and not really sort of validate them going after. So pretty, pretty important piece here. Okay, so a little bit of a, how can we make this very digestible and very simple for clients? Um, today, as you’re thinking about deploying this technology, first of all, we call this the greatest hits.
I didn’t coin that term within the LiveVox organization, although I also would love to take credit for that, but identify your greatest hits, the five things that you think your customers are going to do, and you have a high probability of executing. Um, this is also what I think a lot about, and I give clients all the recommendations to do this is to start simple.
Um, this doesn’t have to be the world’s most amazing bot. You don’t have to compete with Amazon or your lovely airline that you fly, right? That was everything. Um, this, there are a lot of use cases that can be automated in a number of ways that can provide value, whether it is executing an actual function, like a password reset.
Or if it’s collecting information on every call, um, pre the time that you pass it off to an agent, or even authenticating a customer pre the time that you pass it, pass it over to an agent, or simply understanding the customer’s intent so you can route them more appropriately. Think small, um, and execute very, very well on this.
This will help you sort of to build upon moving forward. Then in addition to that, of course, make sure that you have the right dataset and position this within the organization as an iterative kind of a project. In other words, uh, don’t just say, we’re going to build this once and we’re going to let it go.
Think of it more from the standpoint, we’re going to build this and just like we train our people today. We’re going to train this and optimize this going forward. This is sort of one of the bigger mistakes we see. People think of it as, as sort of a set it and forget it. And that’s clearly not the case, uh, with these, uh, with these tools.
Nick Bandy: [00:39:11] So, great recap. So Boris, for, for folks who are on the call today, who, um, you know, are more toward the beginning of their AI journey, what’s the one key takeaway. What’s, you know, if you got nothing else out of the last, you know, 45 minutes, what’s the one thing?
Boris Grinshpun: [00:39:28] Well, I mean, Nick, can, I, I don’t know if I can have two, but, uh, I will.
All right, thank you. No, I, I think, the way that I think about it, first of all, don’t be afraid to pilot this and try this solution. There should be room for innovation within your organization. Um, because I think the benefit is tremendously there. I think the second piece of this, as you, you’re thinking about getting started, get started with a solid unified data set.
It will make your journey much, much, much, much easier. So identify that sort of, some level of source of truth for your data. Um, it will make the deployment much simpler for your organization.
Nick Bandy: [00:40:10] Great. Thanks. So, Michael, we talked a lot, uh, about a lot today. Um, so from your analyst perspective, um, you know, can you bring it all home for us and kind of one big set of recommendations?
Michael DeSalles: [00:40:22] Well, sure. I’d be happy to share, not only some recommendations, but perhaps a few predictions that I think Nick will have a big impact on overall customer service and customer experience. And so if I start on the left with anticipating future customer and consumer needs, this relates, to me, consumer behavior due to the pandemic, you know, and this one’s really hard to predict as we look at a shift in consumer behavior.
I think a lot of it will depend on government responses, not only from an economic stimulus, uh, perspective. Um, if we consider shutdowns and whether or not there’s a sizable uptick in vaccination and fewer hospitalizations, you know, this all ties into our overall economic recovery. I’ll just close on this to say that there’s a tremendous amount of maybe fear, uncertainty, polarization and maybe confusion.
You know, if you layer on some social justice protests and consumer shortages, the topic was just like what we’re seeing with the oil refinery hack with disruption in delivery, you know, this could really be harmful to the economy and the American standard of living. But if we move to recruiting agents and retaining the right talent, that’s blocking and tackling, that’s the fundamentals, you know?
And I want to talk about maybe later in the broadcast about companies that are doing particular things, to have hundreds of thousands of agents that they have to recruit and retain and make sure that they’re trained correctly. Looking at supporting the evolving nature of work with new business models will, well, here are a couple of examples, you know, if you look at the flexible work models today, crowdsourcing, um, the gig economy, temporary and contract work, these are going to emerge and to me, co-exist with full-time work opportunities.
You know, if we expand that into the area of unified collaboration and communication software, you know, this is going to be integrated with business software and all of that will be used to streamline business processes. This is one, that’s not talked about a lot, but there’s a global demand for advanced cyber security and there’s shortage of skilled cyber workforce, uh, individuals.
And I think this is going to create some new security business models, um, that will include threat sharing for the evolving, connected workspace. If we look at the advancement and adoption of innovative technologies, one of the things that we predict is that in a few years,the global cloud computing market will advance to 390 billion with an annual growth rate of 17%.
You know, as companies move toward open-source cloud services that increasingly support in-house remote workforces post COVID, whenever that might be. And you’re going to see a lot of augmented and virtual reality, artificial intelligence and robotics platforms that I think will play a key role in the evolution of the workforce and with that you’re going to see some brand new job opportunities along the way.
I think they’re going to be automation trends, cognitive computing, and the rise of professional service robots that will also transform the future of workplace technology. Because when we look at accelerating workforce transformation, you know, this shift in workforce dominated by maybe Gen Z will come. We predict 75% of the global workforce in 2030, and this Gen Z will need to adapt to the rapidly shifting skills environment in a very fast changing job market.
So that’s going to be a combination of technical and human skills. Now, finally the future workplaces have to seek ways to balance the flexibility offered by contractual work and the lack of benefit and protections that we had in the past for contractual work, by increasing minimum wages and living wage thresholds.
So that’s pretty much all I had then on that note.
Nick Bandy: [00:44:44] Okay, great, good, really good stuff. And thanks for that. Uh, Wrap up on that. So we have some time for Q&A, and a question just came in. That was really good. Boris I’ll direct this to you. What kind of resources does an organization need to help, um, you know, continue to improve their IVA and be focused on that?
Boris Grinshpun: [00:45:03] Yeah, that’s a good, um, that’s a good question. Um, you know, we kind of think about this in two ways. Um, I think one, you must have somebody that perhaps, uh, is sort of, a combination of analyst, slash, customer experience champion, and maybe has had a third hat, which is maybe they’ve worked in a quality management team before.
And the reason I’d say they must have those three hats, um, because they must be sort of looking at it from a lens of, is this experience that we were looking to drive with the virtual agents specifically. Um, to, um, are there some sort of rounding of the corners that we can do and they’re tapping into their prior experience and three, analytically, is this a one-off, is this sort of a particular pattern that we’re seeing other sort of other tangential things, um, that we can round off in a particular request?
In other words, should we approve something that we have? Should we create a new category, how often is this happening? I think those are sort of, like a trifecta of skills that we’re seeking of a person who can actually improve if you will, a virtual agent deployment at a client site.
Nick Bandy: [00:46:17] Great. Good stuff. So, Michael, I’m gonna direct this to you.
You mentioned something a couple of times. Um, and certainly agent attrition has been something the industry has been trying to deal with and it’s becoming perhaps even more interesting or challenging, um, in the new times that we’re in. What do folks do to help keep agents from leaving?
Michael DeSalles: [00:46:40] That’s a, that’s a really great question.
I get that very often, you know, with my lens sometimes on, on the outsourcing industry, one of the things you do is you utilize corporate social responsibility, Nick, you know, that slice, that moves to that fanatical focus on people on their development and their future vocation. And so you want to be known as the employer of choice.
And the question is, well, how do you do that? For companies that are successful at it, you’re doing a lot of different things. One you’re making, you’re making sure that there’s easy access to public transportation. And in some cases you even provide transportation for your employees for night shift. I see really creative things like meals, subsidies, uh, cafes, canteen facilities within the company.
Agents have easy access. And I see this in foreign countries a lot, uh, in this case in Honduras, where you have banking, groceries, post office, all located within walking distance of, let’s say a contact center in a country like Guyana, where you’ve got, um, really serious healthcare issues, particularly for children, there’s milk subsidies, right?
That’s pretty creative and access to medical staff. You’ve even got in some countries, interfaith chapels. I mentioned onsite childcare. And then finally I think that impact sourcing where you’re consciously and intentionally hiring and providing career development opportunities to folks that would otherwise have limited prospects can be a real game changer in bringing in the right people, training them and then keeping that talent.
Nick Bandy: [00:48:23] Great. Good stuff. Thanks. So Boris, a couple of questions just, uh, uh, and you know, back to IVA is a little bit, so we talked a lot about all the things you should think about, to get started and do things, right. What are some of the mistakes you’ve seen made that maybe folks it would help them to avoid?
Boris Grinshpun: [00:48:40] Yeah, absolutely. Um, you know, a couple of, a couple of times sort of come, come to mind. Um, assuming that your business problem is again, I guess like I’ll call it the customer’s business problem. So I think one, uh, what I would advise a lot of people is really to take a customer lens on this, on this thing versus versus not.
Um, the second one is your desire to automate something versus your technical technological capability to actually execute on this. But if there is no service, there’s no API call on your side in order to automate a password reset, this thing isn’t happening in an automated fashion for you.
So I would start to evaluate your own capabilities and then third, you know, um, uh, it kind of goes back to the point that we talked about, which is centralization of data. And I think once we, uh, get into a conversation with lots of clients, you know, the comments are well that data piece is and that data piece is there and that data piece is there.
So, um, I would discourage people from saying, Hey, look, let’s connect this thing to five or six or eight different systems. Then it becomes, you know, a fairly large technical, uh, data movement and service bus, however you want to call it, project. And this is what’s going to stall your implementation. Um, I would advise to centralize data and to execute some of these things in a much simpler, much more intuitive format.
So in other words, you get a high volume, high velocity before you tackle some of the much harder technological challenges about gluing the system to, uh, to everything within your enterprise.
Nick Bandy: [00:50:33] Great then one last question that I see here that came in, um, what’s more helpful to agent efficiency, an agent assisted bot that uses NLP to understand a conversation and make recommendations to the agent, or having a unified view of data from, uh, all siloed systems. And no D, all of the above, is not an answer.
Boris Grinshpun: [00:50:55] Um, I mean, I could make an attempt. I think that’s a uh, that’s an A or B kind of a point. Yeah. I think it’s a little bit of both to be quite honest with you. Um, but I think optimizing and having all of the right data is, is probably more important than, than having, uh, recommendations.
I think that gets you, uh, gets you a lot of the way there. But I’m not going to say that I have the right answer. There’s so many different organizations for, or whose challenges are very different. What I would encourage somebody that’s out there listening right now is I would encourage a test. Um, and, and, and there’s, you know, multiple ways to execute that.
Nick Bandy: [00:51:42] Good. Good stuff. Okay. Hey guys, um, I always like to make sure we give a few minutes back to folks before their next meeting. So great discussion today, ym, before we go, I do want to point out that we have some terrific resources that are relative to today’s conversation that we’ll be sending to each of you.
Um, so Michaelm Boris, I want to thank you both again for joining today and, uh, having, uh, an entertaining and lively discussion. And I want to thank all of you for joining us today. Thanks very much and have a terrific day.
Boris Grinshpun: [00:52:11] Thanks Nick. Thanks Michael. Take care.