How Voice-Based Customer Service is Transforming the Way We Support Our Customers

We have two special guests today from Prime TSR. We have Wally Walikainen, a Solutions Architect, and Rich Kopeikin, an Enterprise Solutions Architect. Today, we’re talking about a growing subject, which is voice-based customer service. There’s been a growing trend of consumers putting voice devices like Alexa in their homes. There’s been increasing usage of this for customer service and interacting with brands. I wanted to talk to both of you today, so welcome to this podcast. I want to get some of your insights on this. I want to start with a very basic question. Feel free; this is an informal conversation. Feel free to chime in. All of us will talk together about this. Let’s start with the basics. What is voice-based customer service? How can it be used?

Yeah, yeah, I could take a shot, just at the high level of this one. Most of us are actually very, very familiar already with voice-based customer service. It’s been around for a long time. Imagine the scenario when, excuse me, you’re on any call waiting or phone answering service that gives you an options menu. “Hey, say yes. Say no. Say which department you want”. It’s helping you as a customer automatically without having to touch any buttons; using that easy voice command, especially while talking on the phone, can do that. That’s the base case that we’re all familiar with. Still, certainly, it’s become more popular nowadays in other areas such as hotel rooms. There are many voice commands for customer service. “Hey, I want to order room service, this particular meal”. It’s become pretty popular. I think we’ve talked about the patient and their experience. I’m getting a little ahead of myself, but that is the gist of what it is. It’s anytime you’re talking to an automated system that’s detecting your language. Based on semantic analysis, it’s determining what you want and responding, or essentially doing a command, doing your bidding.

Wow. Wally, do you have any insights on that? Maybe some examples where you’ve seen voice-based customer service work well?

Many people bump into and maybe don’t think about the checkout lines at cash registers when checking yourself out. That machine is talking to you directly. It’s having a conversation with you. Whether you’re verbalizing it or not, it is speaking to you and telling you to take the thing back out of the area that you wanted it to be in. You scan it again. That sort of thing. It falls into this category where a machine understands what processes you’re taking, has a script, and is speaking back to you.

I didn’t even think about it for that area, as well. It’s so odd that we tend to... We all use it from time to time, but we don’t even consider it now because it’s normal. That makes me think of how McDonald’s uses those touch screens now for everyone to self-order and self-service. In a lot of those cases, obviously, people are just dialing. Sometimes, it helps to have that voice tell you what you’ve selected, or in handicap situations, visibly disabled, or elderly, having that additional option there is certainly a good thing for many people. Sometimes even me. I’m just lazy. I don’t want to have to look at the menu and find a button. Can I order a cheeseburger, please, with extra bacon and more bacon?

I like that — perfect, perfect use case.

Rich, I like that you brought up ADA there. Compliance is an important factor in a lot of this and granting technology so that people have greater access to these things. Absolutely wonderful. It is a great way to use computers to make the world a more accessible place. Sure.

That’s interesting. As far as even with ADA and some things like, what problems are this solving? Obviously, some misconceptions are still there, or some people say, “Well, it’s not a real thing. It’s a fad”. Where would this be used? What problems would this solve where it says, “This would be a perfect solution for what I’m looking for, for what our company is looking for”?

Yeah. I think the healthcare space; we have talked a little bit about it before, is a great area where this could potentially be used — a simple situation. You have a patient who’s in bed. Maybe they have had surgery or, for whatever, are just bedridden at the time. Normally, they might be given a button or something to call for service. That’s about the limit of their ability to get somebody other than the hooked up machines, which will alert certain people if anything’s looking real bad. If you put a device like Alexa in their room, and they could say, “Hey, I’m hungry. Give me some food”, or “I'm not feeling well", or this or that, maybe those type of commands can be translated and delivered to the correct department without a nurse having to take their time, come in, understand what the person needs, go back, then put in the request and contact somebody else, putting in essentially a service ticket, a self-service ticket that gets routed is a great, if nothing else, process-saving and time-saving piece.

It's definitely an effort to remove certain unnecessary people components to some of these things and streamline those processes. Not all. They're always going to need people involved somehow, but to at least take care of some of the easier to remediate tasks. Back to the bagging process before. You don't have to have everybody over your shoulder, checking what goes into your bag — each register you've now taken over. You've gone to a waiting solution.

When doesn't it work well? What are some, either situation, use cases where you said this is not the way it should be? It's more of a novelty. You don't have to have an answer for that, but is there anything that comes to mind where it's like, "No, this is not the way to do it, or it's not designed to answer these complex questions or things like that"?

I think classically, you can identify with the person seeking customer service via a phone call. There's a series of buttons that you have to press to reach the particular solution, the unique... Basically typing, doing a martial Mortal Kombat finger press to get to the right person at the end of the line, versus screaming into your phone, "Agent", to move past that frustration solution. I think we are at a time where that is old. A new version is coming to replace those frustrating mechanics, those frustrating operations with more recent, better fixes for that issue. You want technology to solve a problem and not create a problem between the customer and the service.

What comes to mind is they found that it doesn't really work great in public spaces, essentially. You're outside. You don't want to be listening and shouting into your phone, or in the theater, or something like that where other people are around, and there's background noise. It's just not the most... An office space, even. Nobody is ever going to have these on their desk because that would mean everybody is talking to their thing every day saying what they need. "Hey, I need some groceries, Alexa. Put them in my Amazon cart and give them to me". It would just be chaos. I think the best situations for these are generally the home or some other area where there's direct privacy. It would only be you in the room, or maybe a limited number of people. Again, the ideal situation is for that. Still, there's also, as Wally said, you don't want to overcomplicate a process, as well. When at a computer, I can definitely say that we could all use voice to type. That has existed for dozens of years. Really, not many people use it because you tend to be faster at typing with your hands, honestly. You're faster at clicking buttons and commands and doing things like that. These things work better for specific situations. It is a sweet spot, for sure, finding what situations it does and doesn't work in.

Yeah. I thought I was the only one that screamed at the agent. My thing is that I spam the zero button, the zero on the keypad. There's another trick to it, actually. If the wait is too long, sometimes, if you speak Spanish, it'll direct you to... They're bilingual, also, so they'll get through faster. That's just a trick I've picked up.

I have the same trick. I want to pass over back to you. You're just a domestic flight inside the United States, and the United line is all tied up. You call the international number, the same people on the other end. Still, their international isn't as goofed up usually, because they have to answer more complicated problems. That's just another freebie workaround there. Call the international line, skip the domestic lines.

I'm going to use that. I probably won't be able to use that for another year or so.

This is just helpful. Thanks for all the help, guys. Yeah, when are we going to fly again?

Right. Exactly.

Yeah.

One interesting thing we talked about was the use case and something like that. How do we make this real? Is this a standalone thing? Is it integrated with the company systems? It sounds like it's a one-off thing, but how would a company use it within their overall workflow process? Is it tied in? Is there data? How do we react to that? How do we set that up where it makes sense?

Yes. I think these are mostly going to be... The applications themselves are going to be B2C applications. Consumers will be essentially downloading or somehow using an app that's been created by the company for that consumer. Sure, sometimes, there might be some B2B situations. I can imagine a warehouse or assembly line or something; those workers might be using it to speak in and say things. Obviously, consuming the data from some essential microphone device, no matter where that device lives, as long as it can hear and listen, you're getting those commands in. You want to. Obviously, any company that designs one of these immediately wants to capture every single thing that you're ever and possibly saying. There's been a lot of suspicion and conspiracy theories about that in the past. Beyond that, you really do, from the user experience, when the app is being activated by voice to be used, you want to capture how it's being used, and then target it better for your customers and consumers about, "Hey, this specific feature functionality is possible". It's very much like Google Analytics, I assume, for a website. Think of it that way. Everyone loves to capture Google Analytics for their websites because they know which features and functions are hot. You want to do the same thing for your Alexa or other devices, capture all of that. Wally, are there any others that you can think of data-wise, big data processing?

The importance of that network, of streaming that data to a location where you can then make it actionable. Really important, securing that, getting these devices, these microphones, extensive attack surface. Think of it from a security angle. It could be concerning to have things existing that you maybe didn't create; it was built in a home and now transmitting data into some repository you're responsible for. Things like that. There's a real attack surface there. There can be a conversation over this being used in the wrong way. For the realness of this, I have a niece who's six years old and a nephew who's three years old, and they are both very capable of informing Alexa of things to and not to do. We said earlier that you could do commands quicker on a keyboard and interface better with computers that way. Still, kids are now growing up with this technology in their homes. It's going to be used more and more. Yeah. They know how to use Alexa before they know how to use a keyboard.

That's so interesting. Even deep with this, let's say that I already have a customer experience platform like I already have a customer support platform. Where does that fit in? Do I just plop it in, or are there certain technologies? Also, is it a lengthy thing? Can I do it in a week, or is it six months? What does it take to integrate something like this? How would someone approach that? How would a company approach that?

Yeah, yeah. Really quick, I think this ties into the last question really well. You said it. How does it tie into your customer platform, your whole customer experience, that concept of customer 360o nowadays, where you want to get a view of the customer from all of the different angles. Your customer is probably also using web interfaces to submit information. They might be dialing and talking to live people to submit information. They may even be showing up at your store to buy things, such as Whole Foods would be a great example. How many people could you tie into this information? Again, this is where it gets, like Wally said, a little bit scary. Amazon now has the ability to see what you're ordering on Amazon, like maybe groceries or retail-wise, and what you're going into Whole Foods to buy, perhaps groceries or retail-wise. They could, for individual people, understand, "Oh, this type of person purchases this thing here, and purchases this thing online". Again, targeting those experiences more, hyping things up more, or even targeted advertising more. You want to, again, bring in that data into your full view of the customer. That would include what your customer is now ordering through Alexa verbally, versus the keyboard, versus going into the store, versus calling somebody on the phone? Get all of those together in the same space, and you could do some exciting things, especially if you start tying AI into it. You can get really scary then. Hopefully, nobody creates Skynet. I'll put it that way. You want to. Obviously, any company that designs one of these immediately wants to capture every single thing that you're ever and possibly saying. There's been a lot of suspicion and conspiracy theories about that in the past. Beyond that, you really do, from the user experience, when the app is being activated by voice to be used, you want to capture how it's being used, and then target it better for your customers and consumers about, "Hey, this specific feature functionality is possible". It's very much like Google Analytics, I assume, for a website. Think of it that way. Everyone loves to capture Google Analytics for their websites because they know which features and functions are hot. You want to do the same thing for your Alexa or other devices, capture all of that. Wally, are there any others that you can think of data-wise, big data processing?

We're on the path.

Yeah, yeah.

Yeah. For me, my most recent Skynet moment was Alexa in my house announcing that, "Hey, you might be out of coffee soon. Would you like me to order more coffee for you"?

Really?

Out loud, to me. She was right, and I didn't like it.

Did you order it?

No. I want to do it myself. The idea that, maybe, we're the last generation that doesn't just allow a machine to know when I've used my coffee enough; it knows what's in my fridge. It knows these things about me. To say that they're not making the decision would be... It's almost silly at this point. They know the right decision to make. They know it better than I do. They know I need my caffeine fix, and they're more than willing to provide me access to that.

I was going to say your kids, like you said, Wally, the last generation. They'll be absolutely fine. Actually, they'll probably be angry at Alexa like, "What, you didn't automatically do this for me"?

If you want an agent to yell at, Alexa is in the house to be yelled at now.

What's funny is that I always tell my friends, say Siri or Alexa or the companies, if they collected the data of how much Alexa and Siri were yelled at... I can't even imagine how many times those poor voice assistants have been yelled at.

Yeah. Not only do they get yelled at, but they also recommend blood pressure medication for when they're stressed out at the particular device. Like, "Hey, we understand that you've been frustrated with your Alexa recently. Would you like to purchase some blood pressure pills? Really help you out. Chill out".

Certainly, the other setting that I didn't even think of that this reminds me of is the car, for sure, when you're driving by yourself, and your hands are on the wheel. That's a great time for voice-activated stuff, and many cars have microphones. You obviously have your phone built-in. That's another one, mainly that I love to yell out. The self-navigation system, for sure, when it's giving me directions somewhere, it's like, "Turn left here". I sometimes like to say, "No, you can't tell me what to do". Sometimes I listen, sometimes I don't. No matter what voice I change it to, I always... It's a backseat driver. Sometimes I just don't want to listen to it.

Do you want to talk about the platform? Amazon is one. That's the one I'm talking about, the customer contact center.

There are several different iterations of it. There's a voice recognition portion to be able to interface with it that way. There's also chatbots that interface into your Slack and your Chime. Some webhooks can start to function that way. It's multi-tentacled, this beast is, yes. Most interesting, and I feel like this is one I want to touch on, is they have Amazon Alexa for business now as an offering. We talked about how strange these things would be in your office space, but now with COVID and the pandemic, our office space is our home. My office has Alexa available to me as an option. That is starting to exist more and more in crossing spaces, hybrid spaces, and having those functionalities exist. Use cases may be lacking at times. Maybe that's a lack of imagination, but tomorrow's generation is going to be interested in telling Alexa to schedule a meeting, I assure you.

That's interesting. I think what would be suitable for this, and maybe this is the direction we're going, at least from the content I've been writing. If I have a customer service platform, say like the one you guys are working on, which's the core of this. Help me to understand this. The benefit of having this is that there are multiple integration points within my regular customer service operation. Before, to add voice operating would be very difficult. It seems like Prime TSR has created a way where you can plug in these voice-activated or chatbots and things into the system. You can still have your typical customer service operations. Still, if you wanted to plug in something else, you can do this fairly easily. Let's talk more about that.

Yeah. A good example of that. Prime TSR has been creating this cost-saving, a cloud cost-savings platform. Essentially this is a platform that allows business users to analyze how they're spending money in the cloud. It suggests savings based on common techniques that, "Hey, chop this off. You're not using this; you're not using this resource. Do this here". What it could also do is set up alerts. Alerts for when you increase over cost-spending thresholds. Nowadays, every alert that we get is sometimes; we'll get an email is step one. If it's really important, they'll send a text message. Sure, if you have something that's super, supercritical, and important that you want to know, maybe having Alexa tell you, "Hey, you might want to look at this" because if you don't have your phone on you or you're not in your emails or anything like that. Even if you do, the next step on the phone is the emergency broadcast system. The government gets control of that, but right now, they don't exactly allow everyday people to start making their phone talk to them out of nowhere, but Alexa could potentially do that. It's an outbound alerting system, as well. It's a push notification system.

A lot of that notification happens in text, and I think this next level that we're talking about, this Alexa level relies on the ability to transform that text into voice into audio. Again, for your ADA solutions, one great service that AWS provides now is Amazon Poly. It gives you the ability to take that text, whatever that is; however, you've written it and transform it into a verbal sounding, better than some of the machine sound voices of old. I think everybody's played around with synthesizing speech from time to time. This is now a service that's built into the console, into the AWS infrastructure.

One thing that's actually interesting the more you talk, I honestly approach this more of... When I think about customer service or voice-based customer service, to me, it's like the way I was thinking now, and I've changed my mind. One was I called a customer service agent, and I'm interacting via voice, et cetera. It sounds like listen; we can be smarter in using these voice channels for customer service. Customer service, for the most part, has been very reactive. You call me when you have a problem. We can help you around when you have them or before you have them and give this notification, et cetera, for a very specific use case. That is customer service. It's more proactive versus reactive. It seems like that's where both of you are seeing this to be a use case.

Mm-hmm (affirmative).

Yeah. Ideally, it goes towards conversational, where there's a nuance in the conversation that will help you make some of these decisions. We've noticed that you have run out of coffee, but maybe you don't need so much coffee in your life, could be a form of a conversation where it's back and forth. Yeah, maybe I'll cut back to decaf, is definitely where it's likely going. There are services already in AWS for solving some of that.

Now that works in a hospital situation or healthcare... These are all compliant. How would that fit into an environment like that? Or in a case, like healthcare-related.

It goes back to that; you're really opening your attack surface. The security of connections between your end-user and database, and your logic on the back is paramount. It's got to be something... Security is not a thing that you do once, and you’re finished. It's a constant update to ensure that you're aware of the threats out there, and you're proactively shutting them down.

How do you measure the success of this? How do I know this, if I'm a company and I put in... What should my success be gauged on?

I think we know what failure looks like. It's us yelling at the agent on our phone and hitting zero. We know when it doesn't function. We know we don't like it. We have that reference. I think success is, you don't notice it. You don't want to be thinking about your technology when you're using it. You want it to be ignored like special effects in a movie. You don't want to be, "That's a really good special effect". You want to be saying, "That's an amazing scene, and it's a piece of art that I like enjoying". You want to work your technology into your customer experience where it's more forgettable, and that's going to be a different thing from my generation to younger generations of, "Well, I've been talking to Alexa about this for a long time", versus Alexa entering into our lexicon.

Right. Yeah, that's interesting. It's more of these voice assistant bots, they're not going to replace humans, but they will fit into places where a human couldn't possibly do the proactive nature of these things. I think you could also probably judge success based on the tone of voice. I'm sure there's a lot of analytics that measure... I'm sure if there was like, "How often are Alexa and Siri being screamed at based on these"?

An idea, a solution would be a solution where it's solving its own problems. Be like, "Wow, Alexa has frustrated that person and needs to do a better job and have a better approach next time", and it starts solving its own issues in that way. That does get scary. That should worry you at some point if that talent does begin to exist inside there. I'm not saying it's sooner rather than later.

This is a great conversation that just gets my mind going on all of the possibilities, both good and bad. As you were talking about healthcare, I was thinking of a patient in bed. Maybe they want... Obviously, changing the channel is something they could do, but even playing music that plays itself. What type of music do you like? Maybe what kind of time during the day it likes? Maybe even you're wearing your wearable watch or something, and all of a sudden, the system notices that you're not going to sleep at a regular time. It's like, "Maybe you want to either turn off that TV now, and maybe I could play you some gentle music to soothe you to sleep". Then it gets smarter and knows which songs work, and things like that. It's just crazy what you could think that it could possibly do.

How would a company get started? What is your recommendation for getting started? One approach is to say, "Hey, let's just try to put in some voice assistant bots". There might be another approach of, "Let's just look at overall strategy and journey". How would you recommend an organization or an executive to look into the real use cases of this?

It would definitely be contacting people who have been down some of this journey before. At least the bot’s portion of that. People who have worked within the AWS infrastructure and just know what the process would be like to start spinning up these services. It can be easy. It can have a cost associated with that. You want to be careful when you're doing proof of concepts to recognize that you want to time box some of the solutions to see how well they've worked. I think we have got to be honest about testing things out; your test subject will be human at the end of the day. It's important to get some humans involved throughout the whole process. Yeah.

It's an interesting situation that's similar to software development. Essentially, at the end of the day, it is software development, but I think of it... compare it to applications or web application development, and what's become very popular these days for doing that is an agile process. You deliver the minimum amount of functionality. They see how they use it, and they go on from there. You usually have a BA or product owner who has this vision or an idea of how this is going to work. I think one of the key differences, at least up front, is to have that... A BA nowadays is somebody good at analyzing the business problem, seeing how technology can solve that business problem, and knowing what's out there to do it. Also, they have a lot of experience and familiarity with the tools to solve that problem. In this space, I would ideally want a BA or product owner who has that background, and the experience of understanding what problems can be solved using this that has solved problems in the future, and knows what they need from a functional standpoint. Of course, they're going to... I think an agile approach is still definitely the way to go. Let's start with the minimum amount of what we could do and then expand on it from there. Web application development is hard to get good BAs already. Still, we've been doing it for so long that there are many good ones out there. Who's a good BA for an Alexa project? That's probably a tall order nowadays if you want to find a good one. At least a professional one, because as I found with the user interface design, if you put it in somebody's hands who has that professional experience on it, that's the best way to go.

Alexa, find me a good project manager to develop applications for you.

ever listen to, you know who Naval is? Naval is this tech... He's the founder of AngelList and a few other things. Very, very, very smart. He's a tech philosopher. There's a Joe Rogan podcast, which is about two hours. He says one of the things is that there will be on-demand for high-quality creative workers in the future. Similar to Uber, you get in your car; you press a button; in the future, it could be like, "I need a strong project manager for three months". It's just a notification on your phone. You either accept or deny it, and then you move on. That question could happen. The other thing needs to happen; first, that's just the voice, which is an add-on. That's not the solution. That's more of a way of communicating it.

Yeah. Yep, yep.

We might see that.

More of getting rid of the middle man. Be careful. You're not making recruiters happy right now.

If recruiters were smart, they would develop it themselves, and then they could keep it a secret.

Absolutely. No, it reminds me, I've used it. I don't know if you are familiar with the website Fiverr. I've used it a few….

I use it all the time.

Yeah, yeah, it's great. Those services that, "Hey, I need a video designed, or a graphic designed, or something of that nature, or even just a proofread, a long document, or something like that". You could go up to the $10,000 model, and you could go down to pay somebody $25 to do this, and they're able, and it works out great — same thing. You could access that through Alexa, obviously, as well.

Is there anything that I didn't discuss, or you think it is worth talking about or mentioning, or think it's very important for people considering this? It might also be useful around the technology, like how would this integrate with other things? That's just an idea. If there's anything else that you think needs to be discussed...

There may be a reliability factor, everything from your hardware to your networking to your databases and your endpoints. Suppose you're going to start relying on technology more and more, particularly for vulnerable populations or people in healthcare. In that case, you have to be very strong and sensitive to how you're treating that data. As soon as it starts coming out of their mouth, it's maybe not going to come out twice. There is an importance for ingesting that data in real-time and having a real good uptime for the entire process. Yeah.

I was trying to think of this a little bit more from the development perspective, from the developer's perspective about how you do this. One of the things that I, at least, want to get out of this podcast now, so I could claim credit for the idea when it becomes real, is the ecosystem of WordPress for Alexa. There are many different skill levels of developers and development needs, essentially, just like on web pages. You could have web pages that, in the backend, need AI and severe business processing. You're going to have to hire some pretty skillful people to do that. There are also web pages that are just informational, that need to look great for a restaurant or something else. You should be targeting people who can build those. They should use technologies that are the easiest, like WordPress or other front end generating technologies. Similarly, I believe you could have a really smart chatbots for these apps. Still, you also might just have a convenience app out of the box like, "Hey, if you have a store, program this thing for any store at all". Here's your product, and somebody could talk to it and put it in. It's generic. WordPress for Alexa, that was my idea. Patent pending.

Really good, Rich.

The reason Rich doesn't work for Prime TSR anymore is documented.

There are times when technologies show up like WordPress, and why it never got replaced. It's got such significant longevity; it is really interesting. I think trying to be the next technology is very exciting. Something that fits everyone; everybody fits into WordPress, is very impressive.

It's also happening with Shopify. WordPress can do everything Shopify can do. Everything. That's all. They have all of the integration with Stripe. It's just interesting to see that the solution's usability and very specificness become a platform that the whole... A lot of people are building on.

Absolutely.

This broadcast was brought to you by WordPress. WordPress, 50% off [crosstalk 00:34:34].

I don't think we've stayed loyal to anyone in this….

Use promo code PRIMETSR. WordPress, it works for everything. Except for voice. That you need...

They're getting there. I'm sure it'll be Alexa plus WordPress for web pages, not just for the apps, as well. Maybe it was just my younger self or only a developer web application as a younger developer. I was always like, "Oh, WordPress, that's just the weak stuff for those", or whatever. Now I see how powerful it is and what you can do with it. Even myself, when I'm like, "Hey, I want to develop my own little web page to blog or get information out there for people". I'm like, "Even though I know how to create one from scratch and do all of this stuff, why wouldn't I want to use WordPress"? It's great; it's simple, it's easy. It makes it look better than I could do with my own UX design or UI. It sold me, as well. Again, Wally, that promo code.

The promo code is PRIMETSR for 50% off WordPress.

Hey, Tal, Eugene, we got a new revenue stream. We just discovered it. It's going to start a little bit slow, but we're going to build it up. Hey, Eugene, this is adding value.

Veronika's great at it, and so are Brittney and Stacey. You probably used it in the past; I'm guessing or working with people who have it.

To wrap up, I'm going to ask for some final thoughts. You are recommending to an executive or a technology team. What would you tell them as they begin their journey or think about voice-based customer service?

I think I would tell them to... There are two ways to start, and that's from the top down or the bottom up. That's generally how you look at a lot of these problems. My personal preference is starting from both sides, actually, just taking your time and slowly. Have your hands-on guys do little POCs, get familiar with the technology, so that if you have those big ideas, they could do it. Or themselves, they'll start generating ideas as you work on it. From my web application days, I know that as a developer of a website, I felt like I was the power user who, why wouldn't you want this functionality? It's a great idea. I was the one who knew more about it than most of the users, or the BAs, or something like that because my day revolved around it. Obviously, you want to do the same thing from the top down. You want to start having think tanks, things like that essentially to find out, and your BAs, "Hey, think about how we could start integrating this". You want to crawl before you walk and walk before you run. That would be my approach to an executive, how to get started. Contact your BAs. Maybe get your smart people involved. Again, you dedicate the resources you want to commit, start slow, and go bigger. See where it goes. You have to begin spending somewhere.

That's….

Similarly, thinking that ten years ago, the cloud was not anything as big as it was. It was a great and important adoption for many companies. I think voice-based customer service will definitely be at least a part of what you should be developing. It may not be the fix to all of your business troubles. It may be challenging to implement the first time. Still, I think it should become a part of the conversation that you're having about how you're solving some of your business problems, or creating some business opportunities, as well.

Certainly, I'll amend my statement based on what Wally had been... We play well off each other. You give me great ideas. Just like any other industry, if you want to build something and need it for your business, but you're not an expert in it, reach out to the experts. Just like a marketing agency... We have a business. We know to engage marketing people who have that expertise, or bring them in-house, as well. You want those who have delivered, who have done it, who knows how to guide you, and who knows what works and what doesn't work. Maybe you have that top-level idea. You need your people to develop the ideas and views of how you want to integrate it with your business. Still, those people, then maybe the builders of it, may perhaps even help flesh out their ideas more. You need to work with those consulting companies. Amazon provides those services. There are tons of great Amazon partners out there. Certainly, Prime TSR is climbing that mountain more and more every day with our Lambda specialty. What's the next one you're working on getting us, Wally?

EC2 instances and Windows. Yes. We're working on it.

As soon as I get my Alexa certification and a few other people, we're going to be right there in the thick of it.

50% off AWS if you use this promo code, PRIMETSR.

Yes, I'm sure. I'm sure they're totally giving up those... Yeah. For a margin-based business, 50% is really cutting it close there, Wally.

I'm pretty sure Jim actually has those in his back pocket even. I would not be surprised if, when he goes to clients and says, "I think if you work through us, we might be able to get you a little bit of a deal". He has a couple of those monopoly cards in his pocket. You just inherited 50% off.

We're trying. We're trying.