eCommerce Site Shares Journey to Conversational AI Virtual Agents
Purchasing Power deployed AI-powered virtual agents to automate the top call drivers in their call center. Within the first 3-months, they were automating ¼ of their call center volume that used to be handled by live agents.
Chris Gillen, VP Operations at Purchasing Power, will share their journey toward conversational AI automation and how it has impacted their CX and bottom line.
What You Will Learn:
- Where to start with conversational AI
- How to architect a great human-to-machine CX
- What to expect from a virtual agent deployment
eCommerce Call Center Shares Journey to Conversational AI Virtual Agents
Brian Morin: Well, good afternoon, everyone. My name is Brian Morin. I am the chief marketing officer here at SmartAction. I am your host today for today’s webinar. We do these customer spotlight webinars every couple months, and it just so happens that this month we couldn’t be happier to have a customer of ours, Purchasing Power. From a behavioral standpoint, they act as really an eCommerce platform and the subject today is talking about how Chris has taken their call center and sharing a little bit about their journey to conversational AI and using AI-powered virtual agents to automate many of the call types that had been previously handled by their live agents. Chris, good to have you on today’s call. Can you hear me loud and clear?
Chris Gillen: I can, Brian. Happy to join you today. Thanks for having us.
Brian Morin: Good, fantastic. So just so everyone knows as you’re as … Actually everyone is joining in right now, there is a Q&A box. You’ll see the icon on the top of your screen. Today’s webinar is going to be done in a Q&A format, more or less a free style or free form conversation between Chris and I as I stand here and pepper him with a few questions. But as much we have a preset questions to ask Chris to step us through his journey, we’d love to hear from you and any questions that you have. So as you have questions or comments as they pop up, just use the Q&A box and let us know, and we’ll cover the first … Oh, this may go 20 to 30 minutes between Chris and I, and once we’re finished, we will move to a full Q&A at the bottom of the hour, and we’ll just try to hold on, Chris, as long as we can, to answer as many questions as they are coming through.
Chris, I’ll introduce you in just a minute to talk a little bit about your role and responsibilities at Purchasing Power, and then tell us a little bit about Purchasing Power. But first, for those that have signed in, you might wonder, well, you’ll first … SmartAction, if you haven’t heard of us, who are we? Do we have any street cred in this space? Now, in front of you is the arc overview slide, and frankly I’m not a fan of long overview slides. So let me make this quick. SmartAction started actually as an AI research company back in 2002. So we’re not the new kids on the block when it comes to AI development, machine learning, natural language processing. We’ve been doing it over 15 years, and now support AI-powered self-service across 12 different industries.
We’re now servicing hundreds of different call types and chats and where we fit. This is with companies who already have a contact center platform, or maybe IVR, might be a Genesys, Avaya, NICE inContact, so on. They lean on us to expand self-service automation in their environments either because they lack the capabilities of conversational AI, or it’s just too hard or too costly to build it to their telephony platform, and frankly it doesn’t work very well. So due to our leadership here, Gartner named us a cool vendor of the year. Frost & Sullivan recently named us the leader in AI enhance self service solutions. If you are more interested in what peers have to say, you can see our reviews on Gartner Peer Insights, where Gartner has us listed among handful of leaders, and it just so happens we are both the top rated and most reviewed solution there among the likes of IBM, Watson and others.
So just had to get that in, and just to give you a little bit of our] street cred on the topic. Now, Chris, over to you. You’ve been in the customer service space for quite a while. I know that you were the VP at Toys “R” Us, right in their call center operations before making the move to Purchasing Power. But why don’t you tell us just a little bit about yourself, a little bit about your background and what you’re doing now.
Chris Gillen: Yeah. Hey. Thanks, Brian, and thanks for having us again. I’m really excited to talk about what we’re doing with SmartAction and some of the other work and stuff that’s being done here. But yeah, I’ve been in the retail space for about 25 plus years, worked for a variety of different companies. I did my last stent with Toys “R” Us as the customer service and operations person and did some pretty cool things there I think in terms of just how we were thinking about the evolution of the customer experience. I joined Purchasing Power about two years ago, and a great organization, and we’ve been able to replicate some of those exciting things from Toys “R” us and bring them here, and I know we’ll get into that. Is it okay for me to talk a little bit about Purchasing Power? Just so…
Brian Morin: Yeah, absolutely. Yeah, so please, Chris, tell us just a little bit about Purchasing Power and what Purchasing Power does.
Chris Gillen: Yeah. I’m sure most of you on the phone are saying, “Well, I’ve never heard of it.” I had not either prior to joining. But I’ll give you the elevator pitch if you will. Been in business for about 18 years, Purchasing Power sells into companies as a voluntary benefit, and so we work with a number of Fortune 100 companies, and their employees can actually use our eCommerce platform to shop for goods and services and pay for those via payroll deduction. We typically cater to those employees in your organization that are what we call financially fragile, meaning that maybe they don’t have great credit, they don’t make a lot of money, they don’t have cash on hand and there are purchases or services that they need that the alternative might be high interest rate credit cards or rent to own. So with our program, it’s really kind of a fixed cost.
You come in, you buy the product, there’s no financing charges, there’s no credit check. They can pay over 12 to 18 months typically, and that comes right out of their paycheck. The company’s had massive success over the last 18 years, and I’m sure most of the people on the phone stay up to speed on just the changing nature of the workforce and what employees are looking for and what their needs are. So we’ve had great success really helping people get access to goods and services they need that they might not otherwise have been able to get. So it’s been a great two-year journey here and we’re really doing some cool things I think around just rethinking the customer experience.
Brian Morin: So Chris, before we just jump into that, maybe you can start with just some high level data about your contact center there at a Purchasing Power. If you have facts, figures, call drivers, things on that line.
Chris Gillen: Well, today we actually use two BPOs. I have a small number of people here we’re based in Atlanta, and I have a tier two escalations team here on site that handles some more of the complex issues and ticketing related issues. But primarily, we’re an outsourced organization. We do off shore. Almost all of our chats, which I’ll talk about I think a little bit later, is done off shore. We have a program in the Caribbean and we do the rest of our work out of the Central America. Quite honestly, we sell as a voluntary benefit, but the core of our program is really an e-commerce program. So you come in and you shop the eCommerce platform, just like you would a normal e-com, a Target, a Walmart and Amazon.
So we have the complexities and the same problems that traditional e-retailers do in terms of where’s my order and shipping related issues and all that, all the fun stuff that goes on with just being in the retail space. On top of that though, we also have this financial component because when you actually make your purchase again, it’s payroll deducted, so you’re going to be dealing with us for the next 12 to 18 months. So there are obviously issues and things that can happen there. So, it feels like an e-commerce company, but it also feels a little bit like a FinTech or a financial company as well. The last thing I wanted to point out, Brian, is that we consider ourselves a customer level of effort shop.
About 10 years ago, I got introduced to the book, The Effortless Experience by Matt Dixon, and I would just encourage you if you guys haven’t read that, if anybody on the phone hasn’t read that, I highly encourage you to do that. It really shaped a strategy for me over the last 10 years about how we’re looking at providing service to our customers. As I talk about how we got involved with SmartAction and implemented SmartAction, that’s a key component of our decision making.
Brian Morin: I know a lot of shops, Chris, they might have a bent towards CSAT or maybe an NPS, but for you guys, it’s around the level of effort related to the customer. How did that become such a cornerstone strategy for you?
Chris Gillen: Yeah. If you read the book, I mean, the book really paints this picture that NPS in my opinion is always going to be a metric that is watched and monitored and used in boardrooms for report outs. It is really the health of your brand in my mind, and that’s what-
Brian Morin: Yeah. That’s a good way to put it.
Chris Gillen: But it doesn’t get you there, it doesn’t get to the tangible, actionable things that I think drive NPS, especially in today’s world. I like to look at it as level of effort is how the customer perceives the effort required to solve their problem. So today’s customer, and I think this transcends our business and goes into many different verticals, one, they don’t want to have a problem and I think that’s priority number one is how do you create this seamless experience so the customer doesn’t have any issues?
If they do, then where in a traditional CSAT model, you might say, “Well, we’ve made a mistake and now it’s about surprise and delight to fix the issue,” we say, “Efficiency and speed fix the issue,” because the customer more than anything I think just wants that problem to be resolved in the easiest, most efficient way possible. We model everything we do on that premise. You still have empathy, you’re still certainly sympathetic to the issue of the customer had, but you solved that problem in the most efficient way possible and you find that when you do that, it directly correlates to how they feel about the brand, i.e. driving NPS.
Brian Morin: Yep. Great. That’s a great explanation, Chris. Now, before we jump into how the solution works or your challenges that you were trying to solve and how that led you to implementing conversational AI, I think that the easiest place to level set for the audience is at least to first explain, well, what call types are actually being automated right now so we can put some framework around this. So maybe you can share, Chris, a little bit. You haven’t been in production long, but at least in this initial first phase where you’re at, what call types are you automating with virtual agents?
Chris Gillen: Yeah. It’s a great question and it ties back to what I said earlier, is that the core of our program feels much like e-commerce so you would have … you can deduce that where’s my order and tracking related questions around deliveries certainly drive a healthy portion of our call volume. Things around account balance and spending limits because there are spending limits within the framework of what we do and account balances. Things like contact info, updates, password reset. Those were all things that when we dug into our key call drivers, we’re easily in that 40% to 50% of our overall call volume. We also applied the lens of a level of effort assessment, meaning we actually pulled customers and said, “If you have a tracking related issue, how difficult is it for you to get your answer?”
What we found is that we had the significant percentage of our call volume that was eating up a lot of all of our agent’s time and resources, but also considered by the customer to be this very high level of effort. I had to call in, I had to get an agent on the phone, because prior to SmartAction, all of our calls went directly to a live agent. I had to be authenticated. That takes precious seconds, and then to ask my question of where’s my order, and me to spend two or three minutes on the phone just getting that question answered. That’s a high level of effort. So that’s the framework or lens we used to determine our starting point with what we were going to automate.
Brian Morin: Chris, under the umbrella of order management and account management, obviously there’s a number of different sub components, each of those that their own call type and require their own application for self-service. Under those two umbrellas under order management, account management, Did you just start by handling all of those that you named out and or did you start with pieces of that and then grow into this over time?
Chris Gillen: Yeah, it’s a great question. So when we were building the roadmap, we decided very early on that these were the key things that we wanted to implement. We’re also … We follow agile methodologies here at Purchasing Power, and if you’re in an agile structure, then you know anything you roll out you do in phases and they’re quick, they’re short sprints, right? We took time, stood the program up, we turned on the first, what we call track, and then we let that track set for a week or two, then we turned on the second track. So it’s a very scripted process to how you get from flipping the switch, so to speak, to get all tracks fully live. I think that was a couple of months period of time, three months exactly, from the time we were ready to go to the time we implemented all five tracks up and running.
Brian Morin: Chris, usually the most common question that we get asked from somebody particularly in the area of conversational AI, I mean if somebody already has an IVR and they might be doing some of the most rudimentary, I mean the simplest call type set can be automated with DTMF or Touchstone, they want to start pressing in and be able to automate a little bit more. Let’s start first with your shop, how did you first go about identifying just the right place for you to start with conversational AI?
Chris Gillen: Yeah. Well, I’ll go back to what I said and then I’ll expand on it. The first thing we did is build a roadmap, and I think that’s really important. It was just at a conference last week where AI was the featured topic of the whole conference, and you hear success stories and failures. The one thing I walked away with is, I would tell anybody, you have to have a clear roadmap first, regardless of the tools and technology you’re going to do, is what are you wanting to accomplish? Is it an in-state, if I want to decrease my cost, I want to increase NPS, I want to decrease level of effort? You have to really think about those things. For us it was, we have to reduce the level of effort necessary for our customer to have a great experience.
One of our key metrics here is contact orders. So for every order, how many times am I needing to get in touch with an agent? That’s a big measurement for us. So objective one was how do we lower this whole contact order ratio? So that was step one. Step two was doing a deep analysis on what our call types are and breaking those into chunks, and then smaller buckets to really understand, all right, it’s almost the 80/20 principle, right? You know that 20% or 30% of your calls are driving the most complexity in your call center, they’re taking up the most time. So we looked at that and said, “These are things that we think are easy to automate and are going to reduce the level of effort necessary for the customer.
If the customer has a question about a tracking number where their order is, they don’t need to talk to a person at that stage. We wanted to make it very simple for them to call into to the IVR, ask the question that they need to, have the system pull back the answer as fast and as accurate as possible and allow the customer to disconnect. That was the other piece of it, is just saying, “What are they going to be the easier things to automate?” You have to remember too, for 18 years Purchasing Power never had an intelligent or automated phone system. For 18 years customers have just been calling directly to an agent, and so we also thought about that shock to the system the first time the customer was going to call, and suddenly presented with IVR, this intelligent assistant who can help me with anything.
So we thought about all of those components and said, “This is where we feel like we’re going to get the most bang for our buck, where we feel like the customer is going to feel good about the interaction.” Because I think if you try to do too much complexity right out of the gate, you frustrate the customer, and then the customer says, “Well, I don’t want to call the IVR. I don’t like it,” and then you worry about the customer just continuing contact all together. So those were some of the things that we used in thinking about where do we start.
Brian Morin: Good. Well, Chris, in just a minute, we’re going to talk about how the solution actually works in your environment and talk about some of the results that you’ve seen both from a CX and a bottom line standpoint. But let’s bring up this conversation first, just around the … There’s always some compelling event that drives somebody to consider conversational AI, the need to automate more. Now, I would say … Onscreen, we have the question on the challenges you’re facing. It made you look for virtual agents, but with some of our back and forth while we were talking earlier, Chris, I think I should back this up actually and say what challenges were Purchasing Power … What challenges did they have in the first place when they brought you in, and what were you commissioned to do that ultimately lead to you considering you’re automating more in conversational AI?
Chris Gillen: Absolutely. Yeah, listen, two clear objectives were given when I joined the company, is how do we increase the experience? Again, put that in terms of NPS, how do we drive our overall NPS number up, our CSAT number up, and how do we reduce costs to serve? This is a business that’s been growing double digit growth year over year for a long time. Of course we all know in associate e-commerce call center that your growth comes from orders. When you’re growing your order population, 10%, 20% year over year, if you don’t have the right infrastructure in place lowering that contact order, your costs are going to follow suit. Then eventually, as we know, most service call centers are cost centers, and so that gets very expensive. That was really the charge. It sounds like a big mountain, right? I want you to come in and increase NPS. Oh, and by the way, lower costs too.
Brian Morin: Lower costs too. Right. Yeah. Easy.
Chris Gillen: So that was the call to action for me.
Brian Morin: Okay. Good. One side improve the CX. On the flip side of that is lower costs, and so ultimately-
Chris Gillen: That’s right.
Brian Morin: … then that’s what caused you to look at, I guess it would sound to me, conversational AI as the first phase as attacking that commission.
Chris Gillen: Yeah. But one of the thing is when you’re using the lens of effort, the first thing that you do is you build out an effort map, and how you do that as you look at however you service your customer today, whether it’s phone chat, email, ticketing system, whatever it is, and then you make some assumptions though, and you say, “All right, so what is the lowest level of effort that that can be had in experience today?” Well, obviously that is a seamless I come in, I place an order in our world, I don’t have any issues. Everything works as it is. Then you say, “Well, the next tier is what if I have an issue, what’s the least level of effort necessary to solve that problem?”
Then you start to put it in terms of, all right, is it a call to an agent? Is it a chat? Is it an email? And you begin to model out this kind of no effort to higher level of effort, and then you back into it to say, “Well, how do we begin to reduce this cost to serve the customer? Where can I begin to interject technology or processes that are going to lower the level of effort necessary for the customer to solve their problem?” For us, because of our 18 years, the customer was used to calling, it was natural for us to say, “We need to start with an intelligent system.” Not moving in phases to DTMF at first, to push button.” We said, “That’s just going to be more frustrating to the customer. Let’s move quickly to how we can begin to implement a self-serve intelligent self-serve model via the channel that our customers are most in tune to contact us,” and that was our starting point.
Brian Morin: Chris, I think that right now might be a good time to tee this up for you because when you were looking at, from a strategic standpoint, knowing that you needed to implement some self-service to achieve your objectives, can you just talk from a high level? Any reason why you went with a third party cloud-based company like SmartAction for AI-powered self service automation? Because you could have gone down the more traditional route of an IVR and by the speech rec licenses. You’ll find a third party to build it, you’ll pay for the professional services to program it and so on and so forth. When you were weighing your options, why did you go down this route?
Chris Gillen: Yeah. I think it’s important to point out, we did a full RFP, although I was certainly familiar with SmartAction from a previous life. We wanted to do our due diligence and went down a full RFP road, and we looked at two other vendors as well, and really it came down to two very specific things for us. The first was we wanted intelligence. And I think that when you start looking in this space, and if you do your due diligence, what you will find is a lot of companies may say it’s an intelligent system, but what you find is it’s only as intelligent as what you feed it, meaning that if I’m tying it back to that they cue a knowledge-base or something of that nature, then it’s going to basically just pair it or repeat what I want it to. What we wanted to find was a system that was intuitive enough to connect to all of our other systems, which in our world is not an easy thing to do.
We have very disparate systems here. But something that was easy enough to connect to all of our disparate systems, pull information back, and then more importantly learn from the conversation that the phone system, IVA, as we call it, was having with the customer. Meaning the best example I can give you is as you know, when you have this open platform IVR, and the IVR says, “How can I help you today,” there are a zillion different things that the customer could say. I personally not a fan of the systems that tell me what I can and can’t say. I want something to be open-ended, allow me to say, it completely understand what I’m doing, or what I’m saying, and then action behind that, and then get smarter as those things come in. Quite honestly, that really was a no-brainer compared to the other systems that we looked at is that SmartAction actually did that and did that in a very efficient way.
The other piece, no surprises, I was also charged with lowering cost to serve, and so price put a pretty big part is as well. I could easily show based on what we built the ROI I was going to get associated with the spend. I could not do that as clearly with the other players I looked at as I could with SmartAction. So as I sold this and said, “I need X dollars to fund it, but I’m going to get Y in return,” it was an easy sell and we far exceeded what we signed up for, which is always a good thing.
Brian Morin: It is. Chris, let’s go ahead and just talk just a little bit about how the solution actually works. Maybe the best way is just to walk me through one or two call types, coming in and explaining exactly how the virtual agent is going about handling that, and then how that intermixes with your live agents.
Chris Gillen: Sure. One thing, I’m going to back up. I mentioned the tracks that we implemented, but I forgot a very important part of what we actually did that I think had massive benefit is, in any call center authentication is a big deal. Somebody’s calling in and we estimate that the human element of authentication, if your systems work as they should, can take up to one minute of time for me to ask you the necessary authentication questions. So, part of the other thing we wanted to serve, because remember we have over 300 individual clients that we serve and we have different needs for each of those clients, so it can be getting complex. One of the key things was setting up what we call front door application, meaning can I set up the system to intelligently recognize the phone number you’re calling from?
If that phone number is associated with this same phone number on your account, and then ask you two additional authentication questions via the IVR to have you pass authentication before you do anything else and then tie that in if I need to go to an agent, tie that into a CTI pop on their agent desktop that says, “Brian is fully authenticated.” Shaves 60 seconds off that handle time right out of the gate just not worried about that, and plus provide you greater visibility and assurance that that authentication is happening the way you want it to consistently with every call. So the way it works today … Again, the one thing I think I would mention about working with SmartAction is, especially if you’re a company that has disparate systems, APIs are really, really important. And so being able to do the necessary data dips into the right systems and then pull that data back into the IVR in a way that doesn’t erode the interaction with the intelligent virtual assistant is really important.
SmartAction made that very easy in a situation where it’s not always easy for us on our side to do that. So, you call in today and you’re greeted by an intelligent system, and you will recognize your phone number right out of the gate, so it recognizes the call, it dips into our backend system and says, “Yep, the same number associated with Brian’s account.” It’s going to ask you two additional authentication level questions that you just speak into. It’s going to dip into the system and it’s going to verify those. Once you’ve been fully verified in the system, it opens up and says, “How I can help you?” We did purposely did not put that element of you can say things like X, X, and X, because I believe that’s just unnecessary time. But then from there they go into I have a question about my order.
Again, that information is pulled back on that particular customer, and so it’s smart enough to know if I have one open order that I know that’s the order that you’re calling about, so it’s intuitive enough to say, “Brian, I see you’re calling about this particular order.” Yes. Great. So now I don’t have to tell you exactly what order I’m calling about. That’s just a little bit of how it works. You can have this very nice fluid interaction, very quick, because speed’s important. You don’t want to sit on the phone for 30 or 45 seconds while it’s trying to think, and maybe you’ve been on … some that have a little typing on the keyboard sounds are very annoying. We don’t do any of that. But pulling back very efficiently. I think this is the other piece.
The other piece is there’s a secret in our world that if you want to get cost efficiency, you can just close loop your IVR and then make it really difficult for anybody to go to an agent, right? Hope they dumped out. You don’t have to do that. We don’t do that. We say, “Look, we’re certain enough that these things are going to provide you with the information you need, but if you want to bypass that and go right to a live agent after authentication, you can still do that.” I think that’s a really important piece to allow the customer to not feel trapped, stuck in this digital maze of, well, how do I get to somebody? I think that’s an important piece that we’ve allowed within the system as well.
Brian Morin: Chris, if I were to parrot that back to you, it sounds like the first thing that happens when a caller calls and they’re greeted by virtual agent, they’re authenticated. Once they are authenticated by the system, the virtual agent uses natural language to understand why they’re calling in the first place. Then after the virtual agent extracts intent, if that intent is related to a call type that stays within automation, well, then that call is just kept then. You’re on that cloud network with that virtual agent fully contained, or if it understands the intent is something that a human would handle, well, then it’s simply just transferring the call to the live agent at that point along with the screen popup of gathered data at that point, so they can just pick up where the virtual agent left off.
Chris Gillen: Yeah. It’s a great point because the other thing I love about the system is we have what we call hot words. We’ve only built five tracks, right? In the system. We have plans to build more, but you’re right. If a customer says today they want something that’s not one of those tracks today, it will automatically put you right through to an agent and we record those hot words. Then we have the ability to go back on a regular basis and see, all right, well, what are the customers asking for? What are they saying they need in the IVR? Then we found that as a very effective way for you to say, “All right, I got a lot of people that are asking about this,” and then associate a cost to that obviously, and then build out those future tracks.
Quite honestly, this is where … and I’ll share one with you that came out of this was I also manage our account recovery program, our collections program. So we had not originally built a past due flag. We’re tracking the IVR, and we found that we had some customers that were calling and asking for that on top of trying to figure out and effective way. Now what happens, and this is actually pretty cool because of that, is that we know after that authentication, that you are particularly of a past due customer and we can intelligently put you into the right place. That has had huge positive benefits, both for us from just an experience standpoint because we can move that right to an account recovery person who can then help them quickly and efficiently. It’s obviously saved money in the long run. Hey, there was a question, Brian, I thought I would just answer about.
Brian Morin: Yeah, I saw that. I was just going to tee it up. In fact, I’ll just let our audience know we have just one more question here for Chris to share just the results that they’ve experienced since deploying the system, and then we will move to a full Q&A, if anybody’s wondering where we’re at from a time standpoint. Now I still did ask this question of you, and I think that you see it, Chris, just when you were talking about authentication, just saying, “Hey, how was this recorded for audit purposes?” How do you go back into the and see how customers are actually interacting with you?
Chris Gillen: That’s a great question. I’ll tell you it goes back to when we were selecting. We also, part of our selection criteria was also we wanted a system that would have a proper dashboard for us to interact with, to go back and listen to the calls within the IVR, which we can totally do. I will tell you we really like the SmartAction dashboard. I’ve got somebody using it on a regular basis, who’s going in and auditing and pulling up just how the calls are flowing through the IVR, what the customer is saying in the IVR, and yes, all of that is … the authentication pieces, all recorded from an audit perspective in the IVR or in the dashboard. So we know we can go back and audit and say, “Yep, this was a clean authentication.” We know if there’s issues with authentication. So all of that is there. So if there was ever a question, it is a great audit tool to go back and say, “You were verified, you’re fully authenticated through the IVR.”
Brian Morin: Good. Chris, let’s go ahead and just a close on this one really important question, which is since rolling this out, what are just some of the key observations and results that you’ve seen since you’ve moved at least phase one into production?
Chris Gillen: Yeah. Yeah, I’ve sat on these calls as somebody who’s listening before a variety, and whenever somebody presents statistics, I’m always like, “Yeah, come on. You could’ve have got that.” So what-
Brian Morin: Yeah, right on.
Chris Gillen: What I’m about to present to you really is true, is honest to gosh truth of how good this thing portrayed. I mentioned, we were under heavy scrutiny from our financial department too. You’re committing to this ROI and we’re being reviewed on that every month and it’s exceeded expectations. But let me give you some just basic without talking openly about money piece. Let me give you some that I know you will relate to. Since we launched it, we’ve had about a 20% decrease in cost to serve customers. My budget, we’ve been able to decrease … Really that goes back to our outsourcer spend. We’ve been able to decrease that by 20%. We’re getting about 25% containment, clean containment within the IVR.
That’s really what we modeled the ROI out on, was the whole concept of containment. How many of those calls will self-serve and not flow into a live agent? So that’s been about a 20% cost savings that I was able to give back to the company in a very short period of time. Consequently, we also wanted to measure and make sure we weren’t eroding the experience, and we don’t want, when we monitor comments … We monitor all of our customer feedback, lines of feedback. But since implementation, 13% increase in overall NPS and that’s-
Brian Morin: That’s fantastic.
Chris Gillen: You guys know that NPS is one of the hardest needles to move full transparency. There are some other things mixed in with that, but IVR certainly has played a huge role in cleaning up that experience and making it much better for the customer. Then our big measurement, like I mentioned, metrically on our call center side is our contacts to order. We’ve seen a 35% drop in our contacts to order in the short period of time that we’ve implemented. It really has been probably one of the biggest and best projects we rolled out as a company last year that has had such a big impact on cost, a big impact on NPS. As we go through, we do see comments from customers that are saying, “This is a much better experience today than it was before. I feel like if I have a question, I can get it answered very, very easily.” So it’s been a win all the way around.
Brian Morin: Well, perfect, Chris. Thank you so much. For those that have held on through the duration, just letting you know, we are moving to a full Q&A right now. I have had some questions that have come in that we’re going to field with Chris. If you do have to drop off, just to let you will receive a follow up email from a conducive representative that will contain a link to the presentation that will also provide the on-demand webinar to you after it renders likely tomorrow morning so that way you have that recording to share with stakeholders on your team. Of course, if you prefer engagement immediately, or have any questions that you want answered via one-on-one, you can go on our website, SmartAction.ai, and you can just contact us and we can continue the conversation there.
For those of you that have not yet got your question in, just go ahead and type it into the Q&A box, and we will answer as we go along. Chris, before we go into a couple of these other questions here that have come in, I guess I should, selfishly so, just tee up a little bit about your relationship with SmartAction and just ask you from an open ended standpoint, now that you’ve had this experience with SmartAction, where we’ve done the walkthrough, guiding and shepherding the transformation automation and now that we’ve transitioned from that, go into the ongoing management of the automation. For you, what has been the best part about working with SmartAction?
Chris Gillen: Yeah. I think I love the fact that, one, we have a dedicated person to the account and they work directly with the person that heads up my system infrastructure here. There’s this continual feedback loop on as they’re looking from their end on how the system is performing to collaboration with us, for continual improvement. Again, I’m not a fan of the companies that you buy in, then you get great onboarding and implementation support, and then it’s like a goodbye and you’re left to do it all yourself. The relationship with SmartAction is such that they’re providing us with information that we may not be thinking about, or we may not have look at it that way, and they’re telling us, “Listen, this is what’s happening in the current experience. These are some best practices that other people are deploying. If you implement this, it’s going to cut certain seconds out or it’s going to create a better experience.”
So we’ve been working very closely with them on what I refer to as tuning of the system, getting great feedback on … Are people saying things that don’t resonate with IVA? And then how do we tune those so that she’s getting more efficient. Accents across the US can vary, and sometimes it’s hard to understand. So they’re coming to us proactively with items that we can ingest and continue to make the system stronger. I’ll tell you hands down, I think that’s what we love the most about working with.
Brian Morin: Sure. That’s a good point, Chris. One of those differentiators being it’s not a case where something was built and then turn key and then the keys were thrown to you. It’s a case really of working with a partner who’s now living out the continuous improvement and perpetuity to achieve what you need achieve on your end. As a follow up to that, Chris, there’s a question, what do you do if you want to make changes or updates on how a virtual agent is handling a call type?
Chris Gillen: Yeah. It’s a two pronged process. We have to work on our side, obviously, because remember a large part of what we do is just dipping into our systems to get the technology. But we work directly with our customer success person. Depending on the nature of the change or what it is, we put in a change order, all the parties come together. I can tell you the past due track I discussed, we added fairly quickly. We’re not talking months of development. I’ve been very impressed on … Actually, if we have any hang-ups, unfortunately it’s usually an art side sometimes with the complexity of our systems. But anything that we wanted to do, we’ve done pretty quickly. The other thing that we try to do, which I think is important, is in our mind, we have at least a two year roadmap for what we want to do within the IVR, how we want to expand it, how do we want to grow it?
We share that with our folks at SmartAction so that we’re all clear on the roadmap. So long term, we all know where we’re going, short term we know the work that needs to get done. And even in complex situations where we’re asking maybe for some things that are a little bit outside of the box, we have not experienced one issue or delay with being able to work effectively and efficiently to hit our timelines. Just for reference I guess we deployed IVR real hard date of October 1st and we nailed that October 1st date. Yeah, everybody seems to work together and it’s pretty easy in terms of how you want to handle upgrades and changes.
Brian Morin: Yap. Well, Chris, I haven’t seen any other questions that have come in. So, I think a good closing question for you is what would you say to someone who is considering AI automation for self-service? Any best practices that they should be aware of? Are there things that they should know or be looking for?
Chris Gillen: Yeah. I’ll say definitely one is have your roadmap, understand what you’re looking for. AI to me today is a buzzword, and I think you have to be careful and I’m sure all of you’ve been to conferences and AI is the buzzword.
Brian Morin: Sure.
Chris Gillen: It’s easy to get wrapped up in-
Brian Morin: Everything right now. Absolutely.
Chris Gillen: Yeah. In the sexiness of it, have your roadmap, know what you want to accomplish and have your own definition of when you say artificial intelligence or AI, smart learning, understand what that is and make sure whatever solution you’re looking at is going to not only accomplish what you want to do in the short term, but it’s going to … you’re going to grow into it. To me, this tool is really is a Ferrari and we’re going about 20 miles an hour, but that’s by design. It could go much faster, can do much, but having that roadmap and knowing how you want to implement it, what you want to do, step one, step two, step three … Because AI is new in coming in the business world. It’s also new to the customer and don’t forget about how the customer’s going to react to that and have a plan to ensure you’re going to roll that out appropriately so that you don’t I’ll say unnecessarily impact your customer negatively even though you’re thinking, “This is going to be something that helps.”
Brian Morin: Yeah. Good. Well, Chris, I can’t tell you how much I appreciate you sharing a little bit of your experience taking this journey with us. I know it’s of huge interest to our audience that we chat with. They’re always looking at trying to understand where’s the best place to start you, who are the real world users and what are they experiencing. So, Chris, I can’t tell you how much I appreciate your time. For those who are still signed in and enlisting in, we’re just going to let once this webinar’s over you will get the rendered on-demand version sent back to you so that you can share with stakeholders. Then at the same time, any questions that you have that you would like to take up with us, so in a particular call study that you would like us to start with you and examining your call types and understanding where is the perfect fit for AI automation in your contact center, we’re more than happy to do that and you can make that engagement with us. Chris, I don’t know if you have any other closing remarks, but we’d like just to thank you so much for your time.
Chris Gillen: Thank you so much, everybody. Appreciate it.
Brian Morin: Thank you all for listening in and we’ll talk to you soon. Thank you.
Chris Gillen: Take care. Bye.