Skip to main content

© 2021 Excellerate. All Rights Reserved

Episode 14 | Conversational AI in Retail Banking and Investing

October 8, 2021 | 37 min 09 sec

Podcast Host – Madhura Gaikwad, Excellarate

Podcast Guests

  • Jarrett Miller, Director of North American Partnerships at Soul Machines
  • James Sloan, CIO & EVP, Financial Services at Excellarate

Brief Summary

The retail banking and investing industry is not novice to chatbots and virtual assistants, but Conversational AI promises to completely revolutionize the way these services’ function.

A digital person is an enhanced form of conversational AI that leverages sophisticated technologies such as Natural Language Processing, Emotion Recognition, Deep learning and more to enable intelligent computer-human interactions.

In this podcast episode, we talk to Jarrett Miller, Director of North American Partnerships at Soul Machines about the applications of a digital person in the retail banking and investing space. Our SME, Jim Sloan, talks to Jarrett about how these applications can enhance customer experience and save costs for retail banks and investment firms.


Key Takeaways for you:

  • Definition of Conversational AI and Digital People
  • Technology behind Conversational AI
  • Use Cases for Retail banking and investing
  • Digital people as differentiators for retail banks and investment firms
  • Risks and challenges involved


Madhura Gaikwad (00:08)
Hello everyone. And welcome to yet another episode of zip radio podcasts powered by Excellarate. In this episode, we are discussing conversational AI in Retail Banking and Investing. My co-host in SME for today is Jim Sloan who is the CIO and Executive Vice President of Financial Services at Excellarate. Our guest speaker for today is Jarrett Miller, who is the Director of North American Partnerships at Soul Machines. Soul machines is a world leader in humanizing AI to create digital people. And we are so excited to chat with Jarrett about the applications, conversational AI in the retail banking and investing space. So welcome onboard Jarrett and Jim.

Jim Sloan (00:47)
Thank you very much Madhura for that introduction and welcome Jarrett. So just to set a little bit of the context for the discussion, I mean, we, we know that AI artificial intelligence and machine learning is really considered a key aspect of financial services, applications, and solution these days, you know, we it’s used to streamline and, and optimize processes, you know, ranging from everything between credit decisions and quantitative trading to financial risk management, you know, specific areas where, where we’ve noticed and probably wouldn’t argue, we’ve seen AI, you know, making headways within specifically within wealth management, you know, where advisors are using AI to help manage assets, you know, including in construction, constructing portfolios and optimizing trades, and, and also, with evaluating the risk level of financial investments. We also see AI and, and machine learning, you know, has worked its way into banking and lending from anything between predicting bad loans and calculating credit scores to, to building out credit profiles and risk models, really ultimately, you know, leading to the decision whether to approve or decline a loan.

Jim Sloan (01:57)
We also see it in, in banking lending, you know, in, you know, helping to identify automated fraud monitoring and detection tools. Capital markets is another big area within FinTech where we’re seeing trading systems being calibrated to analyse market conditions and identify new investment opportunities, and then automate the trade planning and execution itself. So, these are all different ways that we’ve seen artificial intelligence and mission machine learning, making its way into the financial sector and, and FinTech specifically. But today we want to discuss some specific use cases around conversational AI. And, uh, we have our guest Jarrett here to talk about that. So, to start out, Jarrett, can you help us understand what conversational AI actually is and, and what Soul machines means by a digital person?

Jarrett Miller (02:44)
Yeah, absolutely. Jim and thanks, thanks for the opportunity to, to be here today. And so, when you look up conversational AI, you’re going to see a variety of different definitions. And those definitions probably going to have a couple of three letter acronyms or two letter acronyms in them. You’re going to see, machine learning, you’re going to see natural language processing or NLP or natural language engines, that sort of thing. At the end of the day, the technology is a collection of capabilities that allow machines to understand, to process and to respond to humans. So, in essence, what that means is we are, applying technology that allows for machines to interact with humans in the ways that we as humans communicate. Right? And so, when you think about a digital person, you can think of a digital person as the face of AI, right? It’s a face and you interact with it. but a soul machine’s digital person is vast more than, than just a face. It’s not a puppet, right? we’ll talk about the capabilities of soul machines, digital people in a, in a second, I’m sure. But so, at the end of the day, the big difference in digital people has to do with their ability to emote and soul machines. Digital people certainly are capable of understanding and showing emotions.

Jim Sloan (04:01)
That’s really interesting Jarrett. So, I imagine there’s, you know, quite extensive technology behind all of this. Can you explain a little bit about how it’s done?

Jarrett Miller (04:08)
Yeah, absolutely. It’d be a pleasure. So, there’s a couple of things you want to know about digital people, but before I talk about the key elements, the underlying, essence of the technology, right, it might be worth mentioning where digital people came from. So, you won’t know the name, Mark Sager, but you probably know Mark’s work if you’ve seen Avatar or King Kong or Spider-Man 2. And you saw those creatures on the screen and thought, wow, that was hyper realistic. That’s Mark’s work. He created that underlying technology and he developed that, and he won a couple of Oscars along the way. Mark connected with the serial entrepreneur and tech business leader, a guy named Greg Cross, and the two of them formed the company Soul Machines. They wanted to bring that capability to everyday life. You know, Mark thought there could be a lot more, there could be digital companions, digital nurses, digital bankers, all these sorts of things in our everyday lives.

Jarrett Miller (04:56)
And so they found a Soul Machines. And so, you can imagine with Mark’s, haven’t seen those films, a couple of things you want to know about digital people. One is they’re hyper realistic, right? The second is that they’re, they all have a digital brain. So, when they started the company, they started a research project, and they went out to create a digital brain and a human operating system. So, every single digital person has a digital brain. So, what that means is, like I said before, they’re not puppets. When they see you smile, they’re going to smile in kind, they’re going to return that smile back. When they see that you’re upset, they’re going to interact with you in an empathetic manner, right? Their face is going to, to show that right. So again, every digital person has a digital brain. They have a human operating system. And because of that, they’re able to autonomously animate. So, they’re able to show reaction to the person they, your program to show. If you’re upset, they show empathy that we don’t have to program for that it’s built into the tech. And then finally, you know, because that they’re empathetically capable. And because of that, you know, they’re, non-judgmental, they, they, and they never have a bad day. So those are some of the key underlying elements of digital people.

Jim Sloan (06:07)
That’s pretty cool. Must be some, pretty sophisticated technology going on behind all that, Jarrett, thanks. Thanks for sharing that. So, I mean, at this point we we’ve talked out how AI and, and machine learning is really has made an impact on the financial services market. And we, we’ve learned a little bit now talking about what conversational AI actually is and you know, what is a digital person and all the work that you guys have been doing at saw machine. So why don’t we segue now into, you know, use cases that we’re conversational AI and digital people might feel will impact retail banking and investing. So, when I think about that, Jarrett, the goal really is to improve and personalize the customer experience really. And one of those expectations would be that let’s say a hyper personalized customer experience would be capable of delivering the right information at the right time, through a conversation with a digital person.

Jim Sloan (06:56)
Some of the specific cases that I can think of you know, again, in the advisory world, right? Say someone’s maybe opening up a, a brokerage account, right. And before they’re actually ready to pull a trigger on opening the account, they might want to learn a little bit about the firm that they’re, you know, going to be open the account with or learn exactly what a brokerage account is. They might be some FAQs that you would traditionally only see a potential client click around on, you know, sort of pre-engagement right. You know, you might also, for some of the brokerage accounts that are out there now, the online services actually enable the client to onboard without too much interaction with the financial advisor, they can onboard themselves almost any, and open up an account, kind of following a test list it’s that, that might be in front of them, right. Where they might have to, as part of the onboarding process, confirm their investment strategy, right. And their risk tolerances, or they might have to sign some disclosures. So, I guess my question is knowing that this can be somewhat mechanical, the process of, of onboarding a client and also the client is trying to learn a little bit maybe where they even get started on that. How easy it would it be to deploy a digital person to handle tests like these?

Jarrett Miller (08:06)
Sure, absolutely. So digital people, again, we mentioned before the face of AI. So, all of the knowledge that an organization has is there, and a lot of organizations will have deployed voice skills, or they have chat bot chatbots, or other kinds of technology that they’ve deployed. Right? So, a digital person can easily access that knowledge that sits in the databases of an organization. So, it’s, it’s pretty straightforward to deploy a digital person to answer the frequently asked questions, right? So, what’s the strategy, or tell me about your policy or what do I need to know? Those sorts of things are very straightforward for digital person, and it’s pretty easy to plug into the knowledge that’s already there. So, when we see customers, they’ll start their journey in a lot of cases by looking for FAQs or looking for ways to start. And they’ll think of this first use case, which has to do with can the digital person answer questions, that’ll relieve the burden to my call center, really straightforward use case as customers leans in and they start to understand, oh, a digital person can very effectively answer questions.

Jarrett Miller (09:08)
They can interact with a customer. And then our customers, right, who are thinking about their customers will understand end that these digital people are having that emotional response, that emotional interaction with people they’ll start to broaden the use case. Right. And they’ll start to understand, well, well, this digital person can ask an answer. It’s not just a, you know, a rote response to tell me about the bank or tell me about the financial institution. There can actually be a conversation, a dialogue interaction that will happen. Right. And so, the use cases tend to expand out of that FAQ into more of an interaction, you know, what do you want to do? What are you looking for? Tell me about your risk tolerance. Tell me about, you know, just different elements that a financial advisor, if you will, would want to know with a customer to direct them towards different products or different solutions that the institution’s offering. So yes, it’s very straightforward to have a digital person answer FAQ and what you’ll see when organizations lean into engagements and deploying digital people, they’ll expand that because of that capability to interact, you know, with a customer.

Jim Sloan (10:15)
That’s awesome, Jarrett. Thanks. And in terms of that expansion, and you had mentioned this a little bit earlier, you touched on it, customer service and call center function for me. And I know we’ve all done this, right. You know, had to call up a service provider. And then we kind of get put in a wait queue. And, at times where asked to, you know, press one, if press two, if, to maybe sort of redirect us to some service potential, in getting our, our problem taken care of, in other cases, we’re just waiting and listening to music, you know, ultimately there’s just some way to sort of query the customer and narrow in on the problem. So, I guess what, what I’m wondering is particularly in the advisory world sometimes, you know, might be a little impatient or even a little nervous as an investor, even a retail investor. And I want to talk to my advisor and maybe I need to be settled down because the market’s acting very kind of volatile and can, can a digital person help here and settle my emotions and, and help me to be a little bit more patient and awake you.

Jarrett Miller (11:08)
Absolutely. I mean, that’s the beauty of the technology. And that’s the thing that folks begin to understand when they engage with a digital person, right? So, the digital people are read being multiple point, assuming that you have your camera turned on and it’s a zoom setting, right. And then we’re interacting with that digital person being there they seeing ours, right. That digital person’s reading the face and understanding is this human that I’m interacting with positive, neutral, negative, are the under the stress right. The digital person can, can actually shift the conversation based on the emotions that it sees on the human’s face. Right. So absolutely digital people can help to settle emotions. Right? So, we have different use cases that are, you know, everything from a, a nurse or a digital healthcare worker, right? So, the world health organization deployed Florence the early days of COVID.

Jarrett Miller (11:58)
We worked with them to deploy Florence, to debunk COVID myths. Right. So that’s one way to sort of demonstrate that the interaction between a digital person and a human does conjure up does bring forth a connection that happens between the digital person and the human. So, it’s a long way for me to say, yes, absolutely a digital person can settle emotions or can connect, but that’s the whole point of digital people. That’s why they have that digital brain is to allow them to process what they’re seeing on the part of the human that they are interacting with. When we talk to different brands, be they in the financial services arena or governments or healthcare. A lot of the use cases that’ll come up early will be in terms of call center, right? And there’ll be a desire to deflect calls to handle or triage if you will.

Jarrett Miller (12:50)
Those calls that happen, that can be easily the answer. Again, it, it sort of speaks to the FAQ issue. I’ve got a variety of questions that are continuously asked by, by customer base. And I want to respond to them in a positive and productive way, but I don’t necessarily have the capability of staffing for that, you know, 24 hours a day, seven days a week with knowledge. So, call deflection is a very straightforward, initial use case that we’ll see on the part of brands, regardless of industry that allows those brands to put a face, not just data, but a face in front of customers and to respond to their needs and to do that empathetically. Right. And so, I know as someone who frequents, shake shack that when shake shack moved in my community to kiosk situation where you had to go in punch in your order, and there was no cashier that I was put off at first, but then I got used to it.

Jarrett Miller (13:44)
And recently I went to an airport where I actually had to talk to a human. I would’ve much rather interfaced when a Kiosk was pretty straightforward. I knew exactly what I was doing. Had there been a digital person on there, it would’ve been even better it because I could have spoken my order instead of knowing which screen to go to. So again, call deflection is something that we see it’s right up there with FAQ information providing. And it’s just a way to leverage a digital person early in the interaction with a customer as part of that customer journey and to respond to them in an empathetic manner.

Jim Sloan (14:16)
Cool. So, there’s one area Jarrett, that, you know, just seems like it would be a great use case. At least I think it would be for interpreting emotion and you know, it’s not just unique to the financial services industry, but it certainly is every once in a while, we get asked, you know, to fill out a client satisfaction survey kind of thing. And it’s, it’s usually online and sometimes you can get them by text. But to me, this just seems like a great opportunity rather than just kind of filling out a Dropbox, you know, to actually talk with a digital person and you know, really let them know how I’m thinking. What are your thoughts there?

Jarrett Miller (14:46)
Oh yeah, it’s excellent. So, it’s actually in a couple different directions. I love this conversation right about empathy and how customers are interacting and, knowing where customers stood. As I mentioned, several times, digital people are reading the faces of the humans that they’re interacting with. And during a conversation, let’s think of a healthcare setting and a healthcare setting, I, if we’re talking to someone and they’re in a great mood, we might take them down one path. They’re not in such a great mood. If they’re under distress or duress, or just in a bad mood, we might take conversation a different direction. Of course, in the healthcare setting, if they’re under extreme duress we absolutely need to get an actual human involved, right. But the digital person can pivot the conversation based on the interaction that it’s getting from the human, the digital person can see on the person’s face, where they are at any point in that conversation now, and that allows for them to move the conversation.

Jarrett Miller (15:36)
Now, if you think of that in terms of past tense. So, after a conversation or after a series of conversations, you’re able to go back and you’re able to determine the empathy quotient to the EQ information of a conversation or of a collection of conversation, right? So, we can go back, and we can look. And this, uh, digital financial advisor was talking to Jim at this point in the conversation, Jim seemed happy, but another point in conversation, Jim was unhappy. Now it’s not actually Jim it’s anonymized, right? We don’t maintain video. We don’t maintain voice. We don’t keep any of that in our systems, but we understand we’re anonymized user or a collection of users. Right? All of my users for this month had the, these general feelings or have gotten confused or upset or unhappy or had some sort of negative emotion. At some point in the conversation, we can take that information and we can learn from it.

Jarrett Miller (16:26)
We can craft the dialogue. We can change our offering. We can change the priority of what we’re using, what information we’re providing to that customer. Right. And we can evolve. We can move that interaction forward based on the emotions we see now what’s even cooler you is take that empathetic knowledge that you’ve gained through a conversation or a series of conversations. And you can combine that out with the learnings that you get from the natural language engines that sit behind the digital people, right? So natural language processing engines give you great analytics in terms of inferring detail about the conversation. So, what was the person confused? Where did the conversation flow, where didn’t it come flow? Where did you have problems? What was stuck? You can combine the EQ with the analysis that you get from the natural language engine. And it’s a really powerful tool to use to evolve your interaction with a customer based on emotional information, as well as based on conversational detail and what you can infer from that.

Jarrett Miller (17:33)
So, absolutely I know of no other technology where you can know how your customer was feeling at any given point in the conversation. There’s great tech out there. Natural language engines are very good at inferring where a customer’s state was survey monkey is great for view into what the customer has told you, they were feeling or feel about your brand, but in the case of soul machines, that interaction with that digital person has a capability to give you very clear, actual insight into the emotions of the human interacting with a digital person. Does that give you some insight there, Jim?

Jim Sloan (18:11)
Yeah. That’s fantastic. Yeah. I just, it’s amazing that you, you know, you guys are implemented, you know, all that sort of emotional diagnosis right. In real time like that. Very, very impressive. I do have a question though, for you before my last use case in the event in a given conversation, let’s say, you know, one of our financial service firms wanted to implement a digital person in terms of, you know, what we talked about the call center situation, or even during the survey. And I guess when I think about people and their money and, and they’re worried about decision or something disruptive happening in the market, that they really feel like they want to talk to someone, like a real person. Right. And they might get, obviously, I don’t want to say impatient, but fed up with, or might want a real person. Right. So, I’d imagine that you can program set up parameters for that digital person to sort of then respond and say, look, you know, hold on for a second, I’ll get my colleague for you. And this would be true in any industry, right. That you can, you know, put in those trigger points where the digital person is smart enough to know, okay. It’s time for me to go get my buddy. Right?

Jarrett Miller (19:17)
Yeah, absolutely. So, we have a couple of concepts in, within digital people, how they’re deployed. And one of those concepts is elegant failure. So, when you’re creating that backstory, you want to create a dialogue for the digital person to it, to use just as you would with any voice skill. It’s just how natural language engines work. Right? So, you’ll, you have the ability to first comprehend different ways that a question is asked, but you also have the ability to fail and to fail elegantly. Right? So, let’s say that someone asked a digital person a question that can answer, you know, they might have just missed her. So, they’ll say, could you repeat that? And then they’ll ask the same question again to say, I didn’t quite catch that one more time. And then on the third time, it’s pretty clear that they can’t respond and say, hey, I’m still learning.

Jarrett Miller (20:02)
This is all relatively new for me. Why don’t I get one of my colleagues to connect with you? Let me connect you with my human co-worker. And then they’ll, they’ll transfer the call. They’ll transfer the session. They’ll find some other way of communicating. So, the ability to fail elegantly, meaning not be able to answer a question, but pivot over to an actual human is built into the natural language engines. And it’s certainly built into the technology. Now, the difference between a really good deployment of a digital person and a great deployment of digital person does have the do with, was there a great compelling backstory created? Does this digital person have a personality? And did you write, did you create the content for that digital person in a way such that it allows that digital person to not be able to respond, but to keep that customer engaged and to get them the information they need. IE through an actual human, so long answer your question. That’s actually built into the tech. It’s something that we do every day, and frankly, you wouldn’t want to have a deployment of a digital person that didn’t have the ability to do that.

Jim Sloan (21:01)
Yeah. And I love how you’re referring to that. Right. Are elegant transfer. Yeah. I think that’s a critical, especially for those in our listening audience that might be considering this type of technology. You know and might have some reservations around that. Right. To know that a digital person can hand off things, you know, you want to drop the Baton. Right. And, you know, because the batons, the person. Right. So, you, you want to make sure that you’re, you’re really careful with your client and there, and the connection that you’ve established and handing it off to a real person if necessary. Yeah. So that’s great. So the last case study, the last use case that I could think of possibly for, and then again, this is not necessarily unique to financial services, but in financial services I might want to, you know, learn a little bit about and different invest strategies out there or the different benchmarks that are out there and types of investments I could make and you know, how I want to might, might want to allocate my funds differently. And so, it’s, it becomes very educational, right? And so, when I think of how much there is to learn in this industry, you know, and the training that might be involved, tell me how like a digital person could help me as a customer or even me as an employee at a firm learn from an educational standpoint, you know, using digital people.

Jarrett Miller (22:10)
Certainly. So, it’s really interesting. The more I work with, we call our fellow employees or soulmates, right. And our, our digital people, fellow workers or digital soulmates, the, the more that I engage with digital people, the more I understand, there’s some things that are digital, people are good at. And there’s some things that digital people may you actually be better at, right. In some instances, than humans, because they’re unlimited in terms of what knowledge they can hold. If you think of an investment scenario right. In the financial arena, there’s a lot of knowledge, right? And even within the financial arena, there’s different types of investment, right? Different vehicles, different laws, different regulations, different rules, different needs for a younger investor versus an older investor. That’s a lot of knowledge for any of us to maintain as actual as humans. Right? Well, again, digital people are the face of AI.

Jarrett Miller (22:59)
All of that knowledge exists out there. If you make that knowledge available, if you have investment strategy A for someone young investment strategy B for someone nearing retirement strategy C for someone with high risk D F G H, if you have all of those types of strategies available, the digital person can understand through simple question back and forth interaction with the human, where that human’s at and then call upon that information that sets out there in the knowledge base, you know, and the data of the organization. So absolutely investment strategies or educational settings of any type are really well suited for digital people because they can mass up, a large amount of knowledge on the back and they can call to it. All they need to do is to understand what’s human, the person that they’re interacting with, you know, needs to know and wants to know, and then they can call that information.

Jarrett Miller (23:53)
And you know, what, if it gets too much, if there’s questions, they can answer again, they can elegantly fail over and have an actual human strategist come and talk that customer through what is needed. Right. So, you know, it brings up another point to answer your question there, but it also that’s great. So, it brings up another, you know, sort of thought that I had, as we’re talking through this and it’s how digital people start in the minds of organizations, brands that use those digital people and then where it evolves to, it’s kind of like how this conversation evolved, right? So, we think the easiest use case is FAQ. I’ve got this information and I want to get it in the hands of my customers. And then, so that sort of evolves to, I have a customer service scenario and I need to staff for that.

Jarrett Miller (24:36)
And it’s really difficult to get financial advisors on the phone at three o’clock in the morning, when someone’s worried about the market, they can’t sleep. They want to call, they want to talk about their portfolio, want to understand any number of things that you can imagine one would want to understand about their money. So, it’s really difficult to staff for that, but “Hey, digital people”, I understand they, I think they can do that because they can interact. And then it sort of evolves to, “hey, digital people” are emotionally responsive so they can actually go farther. And then if you think of how this conversation without us evolve today, it’s very similar to how deployments and, strategies to leverage conversational and AI and digital people evolve. Right? And we kind of ended up here with, can they be a sophisticated analyst? Can they provide digital strategy to humans?

Jarrett Miller (25:21)
And the answer is yes. And one of the things I find fascinating is that journey and some organizations, you know, some sectors of the market lean into this more quickly than others. I’ll give you an example. So, I was called to an airport to help with a problem that that airport had. And that problem was making aged passengers feel comfortable, not under stress, not anxious, not experiencing high degree anxiety while they were in the terminal. And so, the idea was to deploy digital person, the digital person would engage emotionally and would bring a sense of calm to those passengers. It was such a great use case. I love that use case. It worked extremely well, right? No one needed an instruction manual. People of all ages would walk up; they’d talk to Iris at the airport, and they would get the information. They would move all on.

Jarrett Miller (26:07)
It would connect right the airport once they understood you could use Iris in a kiosk, they then thought, before they come to the airport, maybe questions about the parking structure or wheelchair for my aunt or mask policy or any number of things that were relevant. And then they thought, you know, before the airport, okay, that’s great. What about after somebody lost something or somebody wants to give feedback, they can engage with the digital person. So, they went to the before, during and after the experience. And then that evolved once the employees got involved and started giving feedback on what the digital person should know right. To build that knowledge, right? The employees very quickly pivoted to how’s this going to benefit me. Can it help me find my insurance card? Can this digital person help me find the contracts for a part for the jet bridge?

Jarrett Miller (26:55)
Can this digital person help me schedule my time off or do my training? So again, there’s an entire journey that’s happening with a customer. And then there’s also a lot of, you know, need to interact with your own employees that happens on a daily basis. Right? So, I watch every day as I interact with our partners and with our customers, companies that understand digital people for a particular use case, then they expand that to include that emotional interaction and they expand it into different parts of the customer journey. So again, this concept of journey is very relevant. So again, digital people are great for solving problems for meeting use cases, but they’re also really effective of being woven through that entire customer journey.

Jim Sloan (27:39)
I really liked that customer journey and I also liked how you, in the beginning of, of your explanation about the customer journey, we’re talking about how sometime back into the, the financial arena, I might wake up at three o’clock in the morning, but having a hard time, you know, falling asleep. And I’m thinking about my portfolio again, you know, the market can, it’s been relatively stable lately, but there are times of up and down that can just be really unnerving. So, the fact that I can actually get up and make a phone call more, go online and interact with, with a digital person and, and talk through this at a time where my advisor that I’m used to talking to is sound asleep, right? That’s a real appeal to me, you know, as an investment firm, you know, you’re providing a service after hours, but you’re also staffing that at a relatively low cost, right? These to me are always that a firm could differentiate themselves, right? Their brand by leveraging digital people. Can you talk to me a little bit more about the marketing aspect of the digital people and, and how else I might be able to differentiate as a firm?

Jarrett Miller (28:34)
Sure, absolutely. I’d love to talk about leveraging digital people to differentiate. And so, I, back to your previous point, you know, if you imagine yourself calling your financial advisor wanted to at three in the morning, and then you can get the digital person and, and that digital person can meet that informational need that you have. But the other need that they meet is they have that empathetic engagement because you want data, you want information, you want knowledge, but you know what? You also want, you want that calming presence. You want to interact with someone, even if that someone is a digital person, right? So, there’s added value there, which does allow us to segue nicely into, you know, how does a brand differentiate itself and, and what can it expect? Right? So, we have digital bankers, digital nurses, we even have a digital cookie coach.

Jarrett Miller (29:21)
And so Nestle needed to, to be able to interact with its customers but it hard to staff 24 hours a day, seven days a week, chefs that can handle cookie catastrophes, right? So, they deployed Ruth a DH version of the founder, Ruth of Tollhouse. And that digital person was an extremely effective deployment for them. So, you know, I would encourage everybody on this call to, if you want to talk to digital person necessarily you can go to we have a couple there and we have a couple links, you’ll also go to and you can interact with Ruth. And one is you learn about digital people two, your cookies will probably turn out better, right? In terms of how do brands differentiate themselves. They differentiated themselves by having that personal interaction with their customers, right. Couple of numbers that might be relevant, right?

Jarrett Miller (30:13)
So there’s that emotional value that I mentioned before, really hard to get that emotional value with any other technology that’s out there. And frankly, really hard to get that emotional value with humans. A lot of times, because you know, we all have a bad day. Digital person will never have a bad day. They’re always going to be positive. You’re going to write the Corpus so that, they are positive. They’re always going to interact in a productive and positive way. Right? So, in terms of differentiation, so let me give you some numbers. One deployment we had in the first six months, there were 200 million earned media impressions. So that means just basically unpaid media impressions, right? Another deployment we had; it was a digital person that was put out there into the market to represent the brand they saw a 460% increase in the click to buy. It was very, very, very tangible for them, not just connect with their customers, but also, they can see the benefit monetarily, right. We regularly see CSAT scores, bump up 20 points, a 20-point bump a CSAT scores. Pretty significant. One of the numbers I like is net promoter score between 70 and 90. Right. So,

Jim Sloan (31:21)
Ooh, wow. That’s crazy. I’m not a marketing guy, but like I know NPS, like in the 50 range is excellent. So, 90 crazy. Yeah.

Jarrett Miller (31:29)
Yeah, absolutely. And we have all those numbers. We’re certainly happy to share those. We, we always want to talk about what the brand is trying to achieve, how it wants to interact with our customers, but we can certainly give insights into, you know, how to differentiate what can be expected, what the upside it is. But I think, you know, really Jim, it comes down to, at the end of the day, you know, I want to quote Orchid Burson the head of digital from Nestle. She was on a blog with cofounders Greg Cross. And she mentioned that the average interaction between Ruth, the cookie coach and their customers was 13 minutes. And, and the mission of Ruth the cookie coach was to share joy. So, what she said was, and I’m paraphrasing her, but she said the value to our brand of sharing 13 minutes of joy with a customer is extremely high, right? So, I’ll just leave it with that is that there are certainly efficiency benefits. There are certain benefits in terms of, it’s hard to staff for these types of things, with an actual person at three o’clock in the morning there’s a bunch of different benefits. There are different ways to quantify the benefits that you’re seeing, um, that are out there, but brands differentiate with a digital person. And I think it really comes down to the fact that they’re able to connect emotionally and empathetically and engage with their customers, through using a digital person.

Jim Sloan (32:48)
Very cool. Thanks Jarrett. We’ve talked about, you know, some of the technology behind digital people, we’ve talked about some use cases specific to FinTech, we’re talking about marketing and differentiation. So how about challenges and risks? What might be challenging to deploy a digital person or what risks might I be assuming, as a firm, you know, by deploying digital people.

Jarrett Miller (33:11)
Right? So, we are GDPR compliant and there’s in, and I’ve got 20 years security in my background. So, when you say risk, I, my years per up, right. So really the big risk is between a good deployment and a great deployment, right? I mean, the risk is in the limitation of how the digital person is used, was the backstory created well, did I create this compelling interaction? Right? Did I follow the art of digital person creation? Did I am compelling backstory? Does this digital person have a personality? Did I bring that in mind with a persona of the customer that I’m trying to engage all those sorts of things? So, if a brand does all that it needs to do, right, and it either has resources in house or leverages companies, you know, like Excellarate right to properly ideate, strategize to create the conversational design, to do the natural language engineering, to do the advanced engineering, allows the digital person to pull that knowledge. Right. Does it have a nice UI/UX, all those things, right. If an organization either does that in house or relies on Excellarate to do that, the reward is quite high, and the risks are relatively low. Right. Did that answer your question, Jim?

Jim Sloan (34:21)
That’s great. Thank you also for the plug there for Excellarate. Yeah. We, we’re really thrilled to be partnered with soul machines and provide integration services around your digital people. And, you know, obviously Excellarate provides a wealth of solu engineering, FinTech, you know, outside of AI and, and machine learning as well, but been an awesome partnership and, you know, really appreciate working with you and, and learning about digital people. So, I guess as we’re getting toward the end of our conversation here, I mean, just some key concepts for me that like some things that really struck out, you know, that you’ve mentioned, that I didn’t really know prior is that digital people have a brain, like there’s a digital brain involved here. There’s autonomous animation going on. And, and, and even beyond, you know, just the processing of information, the intelligent processing of information back and forth, but even more than that is this whole notion of the EQ, right. That I’m being my emotions are being recognized, right. And that the digital person is empathizing with me as I convey different emotion in what I’m saying. And so that whole empathetic customer experience is something that the saw machines, digital person brings to the table. And obviously the embodiment of all these use cases we talked about and anything from fact to the call center, to education and training, you know, all these things, really, these are just ideal use cases for digital people and, you know, really show their wares. Right.

Jarrett Miller (35:42)
Absolutely. Absolutely. You got it.

Jim Sloan (35:46)
Is there anything I left out in that sort of closing summary that you want to underscore Jarrett?

Jarrett Miller (35:50)
Well, I think he captured it right. And I invite everybody that’s listening to go to to, to go to, to just interact with a digital person and see what that’s all about because the market is evolving in this. And as, as organizations are closing, branches are, are getting farther and farther away from their end customers, right? With their workers, the ability to leverage a digital worker, to leverage a digital person and to empathetically engage across various types of meaning, mediums and across that, you know, full customer journey is really powerful. It’s really compelling. And, you know, we’d love the opportunity to engage with customers and just explore. Right. So, but I think you nailed the Jim. You, you got all the key concepts there, for sure.

Jim Sloan (36:31)
All right. Well Jarrett, thank you so much for taking all this time, join us and explain to us conversational AI and digital people until next time, I’m going to hand this back to Madhura.

Madhura Gaikwad (36:41)
Thank you, Jim. And thanks, Jarrett. We really appreciate you taking the time out of your busy schedules to join us today. Thanks for sharing these insightful thoughts and conversational AI. This conversation has definitely given our audience a lot to think about. Thank you everyone for tuning in. Please do go over the notes for more details on this conversation. For more information on financial technology, product engineering and digital transformation, visit our website, Thank you.

ZipRadio is available on these platforms