Hello everyone and welcome to today's 15-minute live demo session with ICMI with me. Well, I should introduce myself first. I'm Brian Moore and I'm head of sales and marketing here at SmartAction.
Also brought on some help. We have David and Phil that head up solutioning over here at SmartAction. Also a new hire, Toby, who'll jump in and help us with this demo. We said that we're going to showcase for you some of the latest when it comes to conversational AI self-service in the area, both of voice.
and chat. Before we jump into that, you may wonder who is Smart Action. We do AI-powered virtual agents as a service.
And what that means is we deliver the full conversational AI technology stack. It's turnkey. It's omni-channel. All of our customers use the voice self-service portion.
In fact, most new clients coming to us today rely on us for more than just voice, for their digital channels as well, chat and text. But I said we do deliver this as a service. So what that means, instead of just selling software licenses or seats and throwing over the fence and wish you good luck, we bundle our technology with end-to-end CX services.
And that means everything, the design, the build, and even the ongoing operation. There's a lot that goes into the day two and beyond related to optimization and tuning that's critically important to having a good experience for your customers. And so at the end of the day, the best way to think of us is stepping in more as a partner. than a technology provider. We're really working as an extension of your team and working with you to ensure that we're delivering not only the CX, but the ROI that was promised.
So conversations with machines are complex and it really does require experts in their field to do that well. So we'd like to think our approach is working for us. See some of our logos on screen that we manage AI-powered CX for.
We are the top rated. conversational platform as rated by customer reviews and Gardner Pure Insights, a 4.8 out of five. So if you want to know a little bit about us, maybe a good place to start is there just to see what customers have to say about us, as opposed to saying, hearing me talk about ourselves.
But with that said, what we're going to do is we're going to jump into showcasing some different demo experiences. We're going to try to keep this straight to 15 minutes. So Seth, if I start coming up on our time post, make sure that you ping me.
What we want to show here, and all of these are just kind of just short vignettes. None of these are long. Just to show you, and clearly these are live demonstrations, so nothing can go wrong in a live demonstration.
So let's see. We are showcasing some of these, as you'll see, scheduling, vehicle make, model, capture. So be thinking of what you want us to try in that area. Alpha numeric capture strings, ordering a hamburger, channel switching.
Some of these that we're picking here on the voice side. The reason that we're showing these is not because they're easy, it's that we're showing these and doing them because they're hard. And some of these are the hardest captures you'll see.
You're not going to see a lot of organizations even attempt to do an alphanumeric sequence capture, let alone someone do it in a live demonstration. And a lot of this has to do with what we can't get into is what we're doing in the area of pattern matching to compare what was heard against what we're listening for to drive up these really high confidence scores and accuracy. But beyond that, we're going to jump into chat, rich media. and show you the very latest on what's happening in deployments now that are delivering your tap or type experiences with image carousels, image uploads, channel switching, form fields, radio buttons, and much more as you'll see. So let's go ahead and just jump straight in.
We'll show you scheduling and appointment. We do scheduling for a lot of different industries. We schedule deliveries. We schedule service for home repairs, for... for a number of different service outfits.
We do schedule house tours for real estate, service appointments for dealerships and manufacturers for recalls. And this year, we started scheduling golf tee times. But the biggest is actually in healthcare. And what I'm going to show you here is an example for doctor's office and dentists.
So let's go ahead and call in so you can hear that. I'd like to schedule a new appointment. Sure, I schedule your appointment. Would you like to schedule your appointment with Mr. Lee?
Yes. Let me verify. I see the name.
Yes. Next, please state your date of birth. September 3rd, 1979. January 1st, 1973. What is your zip code? 9-0-2-9-2. Thanks for the next appointment at 9 for you?
No. No problem. Let's find a new appointment time, 9 a.m.
to 5 p.m., and you cannot schedule your appointment more than a week in advance. When would you like to schedule your appointment for? Do you have anything on Tuesday afternoon? I have an appointment available 4 p.m. August 17th, 2020. Does that work for you?
No, it doesn't. No problem. Let's find a new appointment time. 9 a.m. to 5 p.m.
and you cannot schedule your appointment more than a week in advance. When would you like to schedule your appointment? appointment for? What times do you have on Wednesday? I have an appointment available August 18th, 2020, 3 p.m.
Does that work for you? Yes, please. Great.
We'll see you at 3 p.m. on August 18th, 2021. Would you like me to send you a text message with your appointment information? Yes.
Should I send the text to your number ending in 8443? Yes. I'll send the text now.
Brian, thank you for scheduling with Smokshin Clinic at 6801 Park Terrace Drive, St. 400, Los Angeles, California. Okay. To see Dr. Lee on Tuesday.
And so right there, you just saw that the text was coming through with the information about the appointment as well as the linkable. Google Maps, click to the office, and then we'll continue to follow up with me afterward so I can confirm, cancel, or reschedule and have those capabilities. So what we're going to do, let's jump over to the next one.
So we're going to do a vehicle make model capture. Capturing vehicle year make and model is something that we do in different areas for different companies. We do it for emergency roadside assistance for AAA.
We do it for insurance claims. And we also do it for auto dealerships. So what we're going to just ask, I think some of you said, hey, find some year make models for us.
Go ahead and throw those in the chat. And let's see what you come up with. And then, Toby, we will have you call in.
And let's do three. We'll do three different vehicles. And so let me just take a quick look here.
All right, so Toby, let's give you some luxury vehicles here. Right, I see someone's posted that we got a 2022 Mercedes C-Class in there, so let's give that a call. Please tell me the year, make, and model of your vehicle, such as 2021 Honda Civic. 2022 Mercedes C-Class.
What is that a 2022 Mercedes C-Class? Yes. Thanks.
I'll text you the parts here. Make a model to show that I can filter them from the stream. One sec. Thanks for using our year.
Make a model demonstration. Have a great day. Cool.
And there we go. Came right through. Okay.
Let's do that again. Toby, let's do one more. And then for a second time, we'll move on. And if anyone wants, if we still have time left and people want to come back to it, we can do more.
Let's try one more. Sounds good. Please tell me the year, make a model of your vehicle, such as 2021 Honda Civic.
2021 Porsche Panamera. Was that a 2021 Porsche Panamera? Yes. Thanks. I'll text you the parse here.
Make a model to show that I can filter them from the stream. One sec. Thanks for using the model demonstration. Have a great day. Thanks, Toby.
And this goes back to the NLU tailoring that has to be done on our side with developers where we're comparing what we heard against what we know that we're listening for. And so even in areas that can have a wide aperture, as long as we can know the year we know the make, we can narrow the aperture down to the model so we can deliver really high accuracy on what that response is going to be in return. So now the next one that we're going to get into is this. Certainly be considered the most complex of the bunch. You know, capturing alphanumeric strings is not something that a lot of voice players do.
We do, I think we do more of this than anybody in the marketplace. A lot of this, again, gets back to this pattern matching approach. We do it for manufacturers when it comes to product registrations and serial numbers and model numbers.
We do it for insurance carriers on policy numbers, Canadian postal codes. I know I'm forgetting one, license plate numbers. So, and this could be a real tough one, but if we know and we can follow like a string and we know that there are patterns with strings, like for instance, if we know that the very first character is a letter and not a number, and we know that it's only one of, for instance, four letters of the alphabet, it means that we can tune that engine to only listen to one of those four. And- as opposed to listening for 26 letters of the alphabet. And that allows this really high accurate experience to come through.
So I'm just going to ask the audience, go ahead and pick any one of these alphanumeric strings. We'll do the same with David. We'll just have him call in. Let's have him just try a couple of them. We've put one to, let's see, one to 110 of just different strings.
So if you just want to chime in some different numbers, David. Why don't you just pick one of the numbers randomly and then we'll have you call in and then we'll have you do pick another one. Sure. How about you just pick a random number? That way, you know, I'm not just going off the script.
OK, sure. Well, let's go with number 52. 52. OK, so I'm going to call in here. What is your ID?
A-D-U-H-6. 6-3-7-7-7. Was that A-V-U-H-6-6-3-7-7-7. Yep. Okay.
David, let's try another one. Let's try one. I don't know if that's any harder, but 84 looks a little longer. Which one? 84. 84, okay.
What is your ID? B-L-M-H-928404. What's that?
B-L-M-H. 9-2-8-4-0-4. Yep. All right.
Thanks, David. Now let's go ahead and jump into the next one of this, which is doing a food order. So let's call in.
Let's order a hamburger. Let's see if it can't get our toppings right. Welcome to the school. I'd like to place an order. Can I have it delivered?
I just want a hamburger. What would you like on your burger? Can I get some onions, pickles, extra ketchup, mayonnaise, and maybe some mustard?
How would you like your burger? I like my hamburgers medium well, thank you. Great. You'd like a medium well, burr, ketchup, onions, and a mayo burger. Your order will be delivered to you in 45 minutes.
See you soon. Okay, so it texted me a copy of my order. Let's go into that. And you can see here, if you're looking, I don't know if you can see the very bottom, but yeah, shows you that it captured the full text string in there.
All right. So we do have a channel switching experience that we're going to showcase, something that we do for metal equipment manufacturers when it comes to authentication. And we do an insurance card upload that captures, with uses OCR extraction to capture that information out there. that insurance card in real time. For the sake of time, I think that maybe we'll skip over this because I don't think we're going to hit our time post.
If anybody wants to, has that use case in your organization, come let us know. We're happy to do that one for you live. So Phil, what we'll do is we're going to show just some of the latest in rich media developments. You can see here on screen, some of these we're going to get into.
We actually have two different. chat demos we can show, but I think that with the time post, we only have time for one. So Phil, I'm going to stop doing my screen share and I'll hand it over to you. Thanks, Brian. Share this screen and we're going to get right started here by clicking the chat icon that's embedded in the webpage.
It's going to ask who's testing just to make sure who I am so that we can get it. into this here. And first, I'm going to choose something on my family plan.
We've obviously got buttons there. We've also got the ability to use typing in a response that it can recognize as well. So it's going to run through some information on how to add a family member. And then I'm going to have to input some information. So it's going to ask me for first name, last name.
I've obviously got some saved data so it's nice and easy if you're on a laptop or a desktop that you know of. I can also upload an image in here, so showing that capability there. I'm not going to upload one right now.
Unless I've got an image on my desktop, I can just drag in that. This isn't a license, but we'll upload that one. Please upload my file.
Perfect. It's got that one. And we're away. So we'll submit that.
And then we wanted to showcase the fact that it's going to upload that into a Salesforce case so it knows I need to add a member to my account. But we'll show that at the end because right now I'm just going to go in and also see about managing my account and see if I can also succeed with this task within the demonstration here. So what is it I need to do when managing my account? First of all, they've got to verify who I am.
So let's enter my email. And we've got this birth date scroll in here. So it makes it as easy as it possibly is. I'm going to pick a random birthday.
No one needs to know when it is unless you want to send me a gift. So we're going to choose in there. Obviously, it would verify if we had information on this customer and verify that that is the correct information. I'm going to say I want to cancel my membership, which is obviously going to prompt, you know, do I really want to cancel this or can we save this in some way? With this, we're saying we can pause the membership, but I'm just going to cancel the membership anyway.
In fact, let's pause it. Let's pause the membership. because that's going to create a case in Salesforce for us.
So we see the case is created there. And what I wanted to demonstrate is the fact that if I drag my Salesforce window in, hopefully you can see this. If I refresh, we should have one on the bottom. There was seven. Now there's eight.
So that pause member account has come up with the details that would be in there for that account. Now I can go on and do some other tasks in here. you know if there's delta pricing i can now link my my delta sky miles account uh to uh to my my account here for clear uh all it's gonna need to do is just verify obviously that that delta account so i'm gonna i'm a diamond member i fly a lot um so they're gonna offer me a free membership here um they just need to verify what my uh what my number is let's go there's my delta number Obviously, again, some verification in the background is going on, and then that's going to create a case here, which is another thing that the agent on the back end can fix. Or if we have the API capabilities, we can obviously do that automatically.
And again, there should be another case there on the bottom you see that's just been created. So that's just a really quick, I know we're on time crunch, but that's a really quick view of kind of what it is we've got going on in the world of chatbots with the rich web chat capabilities. And just to finish off perfectly, Brian's on mute.
Yeah, it's taken me over a year to figure out that mute button. I know we ran a little long, hopefully not too long. If anyone does want to see any of these demonstrations live for your team, let us know. You can visit us at smartaction.ai and we'll pick up the conversation there.
Have a great day, everyone. Thank you.