Omilia deploys conversational AI to improve customer interactions

00:00 [This transcript was auto-generated]
Keith Shaw 0:00
Hi, everyone, welcome to DEMO, the show where companies come in and show us their latest products and they show us how they worked. I'm joined today by John Diatto, he is an enterprise account executive with Omilia. John, welcome to the show.
John Diatto 0:12
Thanks for having me, Keith. Really excited to be here.
Keith Shaw 0:14
So what is Omilia? And can you explain kind of what it is and then what what kind of platform we're gonna be looking at today.
Speaker 1 0:20
So in a year where AI is everywhere everyone's talking about it's kind of like the.com bubble. Omilia is a conversational AI company that has veteran status, tenured 20 years in the business 20 years in AI. So big comparative difference when you're looking at other vendors in the market and we've been solely monastic we focus on omni channel, conversational AI.
Keith Shaw 0:41
Okay, so what do you mean by conversational AI then? Can you explain that?
John Diatto 0:44
Sure. Conversational AI, meaning anytime a human is having an interaction or a conversation with a machine, right? In a synchronous capacity, right, in real time, back and forth, conversational? And we're able to do that across all different channels.
Keith Shaw 1:03
All right, and and who is this designed for within an enterprise space? Or an IT space, like this platform is designed to be worked within an existing platform? Or is it a standalone platform that can then be worked into other software.
Speaker 1 1:17
So everything about our technology is proprietary. So everything you're about to see today, the demos I'm going to show you and whatnot, everything is built in house. So it's a proprietary tech stack, we built our solution from its inception to be as agnostic as possible. So when you're thinking about what horse, you have to hitch this to, we play with a lot of different players in the market.
Keith Shaw 1:36
Okay. And so let's get right into the demo. Because again, the scenario that you've come up with is fantastic. And I love how this works. So explain what you're going to be showing me. And again, this is not just only for fast food restaurants, this is the example we're going to use. But this can be used in any kind of customer service application. Correct?
Speaker 1 1:54
Correct. Yeah. And from our inception, we were we were predominantly within the call center space. But you know, AI has changed a lot. Technology has advanced a lot and as such, we're finding our solution being used outside of that specific space.
Keith Shaw 2:06
And what I like about fast food is everyone has gone to a faster than ever since how bad the connection the communities could be. So let's just jump in and just show me the demo. .
Speaker 1 2:14
Absolutely I'm glad I ate I hope you ate too because this usually gets people hungry. Hi, there. What would you like to order? Hey, yeah, are you guys still serving breakfast or no.
Speaker 2 2:28
No, but you can check out our all day breakfast items on the menu. Let me know when you are ready to order.
Speaker 1 2:34
Well Darn. Okay, then can I get a four piece Nuggets with a Diet Coke light on the ice? Please? Would you like any sauce for your nuggets? Yeah, barbecue. But can we make that a 10 piece?
Unknown Speaker 2:52
All right. Anything else?
Speaker 1 2:54
Yeah. Can I also get a cheeseburger, extra onions side of fries, no pickles, please.
Unknown Speaker 3:04
Okay, that'll be all.
Speaker 1 3:06
Instead of American. Can I get cheddar on that? Okay, and I'm trying to lose some weight. So can I get a salad? Any dressing? Italian Of course. Okay, that will be all. No, another hamburger for my friend Keith here.
Keith Shaw 3:27
No onions.
John Diatto 3:30
All right. And that'll be all Thank you.
Speaker 3 3:36
Okay, if everything looks correct on your screen, the total will be $11.50. Please pull forward to the pay window.
Keith Shaw 3:45
So what I like about this is that it makes adjustments on the fly, correct? Well, if you were looking over on that, on the right side of the screen, you could see that it was making adjustments. But then it also understood, just kind of casual conversation. Right, right. Yeah,
Speaker 1 4:02
Yeah. And really what we tried to do and what we aim to do is not punish inherently human behavior, right? Human beings don't think linearly, people think out of out of turn out a thought we speak the way we infer from context. And that's exactly what we're seeing here. Right. Like, even for example, with the cheeseburger, so cheeseburger, extra onions side of fries, no pickles, please. Right, that side of fries has nothing to do with the order of the hamburger itself, the system is able to infer that's a separate line.
Keith Shaw 4:29
Yeah. So is this going to be deployed or hasn't already been deployed?
Speaker 1 4:33
And it has been testable in several places I cannot mention or I will get slapped on the wrist. Okay.
Keith Shaw 4:39
We'll figure it out. But then this can be applied to other examples, right? You have an example where well, you don't have an example here. But you are saying that this could be deployed in a live chatbot setting or with a live customer service agent, right?
John Diatto 4:55
Yeah. So again, everything about communication is about accuracy, right? If you can't understand what someone is saying, you're gonna have a hard time, you know, servicing that person. Yeah. When it comes to automated call fulfillment, right? What allows us to break past? What is a typical glass ceiling of fulfillment, right. So getting your need address in an automated capacity, we're able to break past that because of this inherent, proprietary tech stack, and a level of accuracy, right? Like even in this demo, you're seeing a pretty much a live configure price quote. And to be honest with you, the latency was a little high today, because we're on a public shark and whatnot. But we can see that these are changing to your point on the fly. And you know, colloquialisms like Diet Coke, light on the ice and this level of service and functionality and complexity and being able to understand a human being is something that we can replicate across all channels. Yeah. And we often
Keith Shaw 5:49
And you're working on adding generative AI functionality to just this existing conversational AI. Correct?
Speaker 1 5:55
Correct. So we are very much of the opinion that there is a high saturation of vendors in the market that are overly reliant on generative AI. Yeah, there's a lot of unsupervised learning and hallucinations that are happening. And you could take your pick of companies that have deployed a to loosely let go generative AI. And what we've done is again, I've mentioned 20 years in the industry, yeah, sanitized. A lot of existing fielded customer intents and utterances. And we've complemented those with generative AI. So expanding the language models, so it's hyper focused and a generative capacity.
Keith Shaw 6:29
Okay, and what what's the pricing on on this platform? Is it is it a standalone platform? Or is it is it an add on to an existing structure infrastructure that you already have? How do you? Is it a subscription model? Is it a one time price? It's got like seven questions?
John Diatto 6:47
A couple questions, what I can say is this without signing an NDA. We are subscription based. Okay. So it's largely usage based. However, that being said, in terms of where we can connect again, it's very, very flexible, like here we are deployed in a drive thru environment where there is no uniformity of the hardware conditions on it, and that and but yeah, again, typically our inception was within the call center. And that's largely you know, where a lot of our bread and butter is financial services, healthcare and insurance, things of that nature. Alright,
Keith Shaw 7:18
For people that want more information, where should they go? We're give me the website. Sure.
Speaker 1 7:23
It's omilia.com or ocp.ai. For those who are a little more technical.
Keith Shaw 7:27
All right, John, thanks for being on the show. And thanks for the demo. Keith, thanks so much.