Is your brand ready to thrive in a world where AI-driven agents become the primary point of contact for your customers?
Today, we’re joined by Gil Rosen, Chief Marketing Officer at Amdocs, a leader in reshaping connectivity and customer experiences through cutting-edge technology. Gil recently spearheaded a global survey in collaboration with McCann Tech and Coleman Parkes, exploring the impact of AI-based agents on brand identity.
With Agentic AI emerging as the next frontier in customer experience and marketing technology, Gil is here to help us unpack how businesses can adapt to this transformative trend and create meaningful connections in the age of AI.
About Gil Rosen
Gil Rosen is chief marketing officer at Amdocs, responsible for managing its global brand, product marketing and customer marketing. In this role, he leads development of Amdocs’ vision, delivery of strategic research insights to customers, as well as the company’s overall go-to-market activities. Gil heads a team of global marketing professionals and is a member of the company’s executive management team. Previous to this, Gil served as the chief marketing and innovation officer at Bezeq, Israel’s largest telecommunication provider.
Transcript
Note: This was AI-generated and only lightly edited
Greg Kihlstrom:
Is your brand ready to thrive in a world where AI driven agents become the primary point of contact for your customers? Today we’re joined by Gil Rosen, Chief Marketing Officer at Amdocs, a leader in reshaping connectivity and customer experiences through cutting edge technology. Gil recently spearheaded a global survey in collaboration with McCann Tech and Coleman Parks, exploring the impact of AI based agents on brand identity. With agentic AI emerging as the next frontier in customer experience and marketing technology, Gil is here to unpack how businesses can adapt to this transformative trend and create meaningful connections in the age of AI. Welcome to the show, Gil. That’s good to be here. Thank you. Yeah, looking forward to talking about this with you. Why don’t we get started with you telling us a little bit more about your background and your role at Amdocs?
Gil Rosen: Great. So let’s start with Amdocs, maybe. So Amdocs is an international company present in over 90 countries. 30,000 employees, give or take. And being responsible for marketing basically means that, um, I have a global responsibility for the go-to market for all the different products and services that we, uh, that we do. We have a very, uh, comprehensive product portfolio. So it’s a lot of products to take to market for different buyers. So it’s a different challenge for each one of the market segments in different territories. Also being responsible for all the brand related activities from events to the brand itself, employee activities, internal communication, and customer marketing, which is basically also known as field marketing, which is basically taking our offerings to the specific customers and supporting our sales teams on the ground. So really encompassing all marketing activities globally. And also one of the things also related to this conversation is the fact that I have a research unit within the marketing division. Where we do two things, we forecast trends that we think will be relevant for our customers. So I have conversations with our customers globally, meeting with. The C suite of all the big telecommunication companies in the world, really telling them what’s going on around the world, but also telling them what’s what’s coming. And the other side of it is basically the conversation we have with our technology teams, because our products and services need to be ready for that future. So it’s a kind of internal strategy and also external thought leadership and trusted advisor kind of role, which is very, very interesting. Before that, I was a chief marketing officer for a telecommunication company, the largest one in Israel. So I was also on the other side. And prior to that, I was in Deutsche Telekom in Germany. Four years, launching many innovation projects from concept to market, so also having end-to-end responsibility. Even to the extent that my last venture was actually spun off Deutsche Telekom, which was a very big achievement within that giant. And I won’t go as far as back as having my own startups, but I’ve been on both sides of the fence from being an entrepreneur all the way to being in a large corporation like I am right now and always find the fun things to do on both sides.
Greg Kihlstrom: Great, great. Definitely looking forward to talking about this topic with you then. So this is something that, you know, we’re headed into the new year here, 2025, and agentic AI certainly is on my radar. And I’ve included on a few lists of my own as one of those things that marketers really need to be paying attention to. Certainly, we’ve been talking about Gen AI for, it feels like, decades now, but at least a few years. But agentic AI and customer experience is something pretty interesting, and like I said, a lot of people need to be paying attention to. Before we dive in a little bit deeper, for those a little less familiar, can you explain what is agentic AI and how does it differ from some of those other AI technologies that a lot of people listening may already be more familiar with?
Gil Rosen: So I think the best way to describe it in the way that people would basically most likely know already is that this is the, let’s call it a chat GPT that talks on behalf of a brand. So it’s a persona. And it’s a persona that’s dedicated to serving the specific brands and it can have conversations and provide information and provide assistance and direction and whatever else is required based on, of course, what the brand does. And interestingly enough, and this is something that I say, which is kind of the oxymoron of where we’re heading is that the more technology oriented marketing and customer engagement we’ll get, the more humanized it will be. So in the past tech. kind of created a buffer between the person and the brand in the sense that you had to interface with buttons, with screens, with, with things that are not a human. And with the gentic AI, the, the nice thing about it is that because it’s actually a person that you can talk to, whether a voice only, or even a manifestation of some, some type of avatar, and you talk in natural language. The engagement is seamless in the sense that you don’t have to press a button or look for a menu or do a search. All these things, if you think about it, are like techie and most people find them hard and it’s not by coincidence. And this is taken from some other, you know, uh, research that we do that, uh, the majority of people actually prefer to call. At least service companies before they try their, uh, their, uh, digital platforms. So still the, I’ll say 60 to 50%, depending on the companies that provide services. They, they, they manage huge care support divisions that basically allow them to, to provide the assistance, support sales to whatever, to, to the, to the millions of customers they serve. And this is a big, not only a cost item, but it’s also from the customer perspective. It’s not ideal. It’s not ideal to wait on the line. It’s not ideal to look for things. And the agentic experience basically means that you’re talking to a human being only. It’s not a real human being. It’s a, it’s a non-human intelligence, but it’s super intelligent.
Greg Kihlstrom: Yeah. So yeah, I mean, a couple things there. I mean, you’re certainly a lot closer to the research, but bringing up the idea that menus and these digital interfaces are not purely natural, I wonder what, is there a cognitive load aspect that I to that, that we’re not even aware, like, we’re so used to using websites and having to hunt and hunt around and find things that we’re not even aware of kind of that, that cognitive load that it puts on us. And yet, it takes a toll in the, in the longer term, you know, and I wonder, you know, along those lines, like, how do you see agentic AI changing the way that brands interact with their customers and kind of the, not only from the way they interact, but maybe even just from the satisfaction end as well.
Gil Rosen: So maybe I’ll, I’ll backtrack to how I got started with it in the sense, which basically sparked the research. So last year, almost a year from now in a mobile wall Congress, which is the biggest, uh, telecommunication event of the year in Barcelona. Uh, we actually had two, uh, agentic. uh, avatars that we trained on everything that there is to know about Amdocs or products or services. And they, they would, they, we had this special pod for them and people could basically ask anything they wanted about Amdocs or products. And they knew more about the company than I would because, because they know everything, right? You train them on a load of data and they just know everything. It was only post the experience when I was watching the people actually interacting with these avatars that I realized that. From my position as a chief marketing officer, I never gave enough thought to how I trained the agents, not only information, but behavioral-wise on the merits of who we are as a brand. So there was a disconnect between the technological capability that I had available to me, which I used in the event. That’s one thing, but another thing which I completely ignore and didn’t, and still doesn’t exist in the marketing domain is a basically best practices or how to approach the agentic experience from the brand perspective. And I realized that a year from now, basically a year from then, which is now this is going to be available to all the brands. And there must be a better way to consider what are the right things to do from a brand perspective, because now you’re actually putting a person in front of our customer and he needs to get, or she needs to get the experience, which I as the brand want her to feel from the brand perspective. Marketing people are not in the conversation right now. That was my biggest problem. It was the technology people launching a technology solution and leaving out the brand considerations. I, as a brand, behave very differently from my competitors, and they behave differently from me. So obviously, between telecommunication companies, it would make sense that the T-Mobile agent is not going to behave the same as the Verizon agent or the AT&T agent or the Vodafone agent. The brand needs to be inserted into the agentic behavior. And that’s something that very few marketing people still realize. So I thought, and this is why McCann, who’s my marketing agency is involved. I called up the CEO. I said, listen, we have this, it’s a new domain we have to investigate. And that basically brought me to where we are today, that we have a much better understanding of two interesting angles, by the way. One is what could customers expect? consumers, but also what the executives of those telecommunication companies, um, expect and many times it’s different things. So that’s another area. Very interesting thing about the, uh, the research findings.
Greg Kihlstrom: Well, yeah, and I want to talk some about the research there because as I mentioned at the top of the show and you just referred, you’ve done some research here on AI-based agents. What were some of the most surprising findings and how should brands think about evolving their identities based on these?
Gil Rosen: Okay. So the, the, everything starts with, and this is like branding one or one, the brand persona is something every brand knows. So every CMO marketing professional in whatever company knows what the brand persona is. So that obviously if the, if you think in a very simple terms, if you call a bank or if you call Disney, obviously the agent will behave very differently, right? One will be super happy. The other one might be like serious and concise. So that’s the brand persona. Then it’s the context. Why are you calling? Is it the support a reason or is it a sales reason? What type of support? The bank, your money is involved or is this technical support? So the context. And then who the customer is, you can imagine if my, my 85 year old dad calls the call center, he might want to talk to someone different than maybe a teenager might want to talk to. Now with agents, with agentic agents, basically you can choose, right? So as a brand, you have to make a decision. And I’ll explain, you have to make a decision who you’re putting in front of the customer. Up until now, the notion was next available agent, because we’re talking about people you were waiting in a line and whoever was now available, you got to talk to. With agents, you can actually decide who will be the persona, male, female, happy, serious, concise, or lengthy. So you have, these are all parameters you have to make a decision on. Furthermore. Do you want to have just one agent so let’s think about apple apple have siri right siri is one agent you can look at the siri as an agent. I decided to have one agent and you might be able to change the language or maybe female male voice but it’s siri siri. And it remains siri so do you want a single brand agent persona that talks to everyone. That’s of a distinct age with a distinct voice or do you say I have a persona that I can tailor to put in front of every single one of my customers, which might be slightly different based on their preferences. Now this is a marketing decision, it’s not a technical decision. The tech people have to give the capability to modify the persona, the characteristics of the persona based on the person in front, based on the context, that this is a marketing decision. Do we want a team? Do we want one person? Do we change it? So this is the marketing conversation that I think is not being had In most brands, which must be because it’s a strategic decision and it impacts a lot of the, you know, a lot of the scenarios your customers will be facing. And I think that’s the most interesting part of the setup. Let’s call it. Then what we found was interesting in the sense of most people prefer a woman’s voice, for instance. Uh, women prefer, uh, I’ll prefer women in general. So women want like 67% of women want the woman’s voice on average is it’s 46 and men is like, or, or male voices only like 30 something. So that’s one aspect of it. The other aspect was age. What’s the perceived age of the, of the voice or of the avatar that you’re talking to. Most people want to talk to an avatar, their age. But in Asia, for instance, we found a very distinct preference for talking to somebody who’s older than you, because in the Asian culture, there is a respect for elderlies, which is higher than average, let’s call it, than what it is in the West. So distinct cultural preferences have to be baked into the Asians. And those are things that we have to consider. Also, we ask people, do you prefer them to be empathetic or concise? And so all these different parameters. And people usually put a very high score on empathy. So it’s important to them. They don’t want to be too concise. They do want a little bit of small talk capability. But it’s definitely something that’s Changes between geography so you have to be aware of who your brand is serving and definitely again it has to map back and this is the important point to who you are as a brand. It’s not it’s not something that can be just. Launched like, like, like, uh, bots were launched, for instance, you know, bots are just like, and many people see agents as the evolution of bots. And that’s another aspect in the research that we found, which is a, which could be a very big mistake to make. We actually, we also included, uh, focus groups on the research. And we, uh, uh, we heard very harsh comments from people regarding bots. So if you want to launch your, your agent, probably it would be a mistake to brand it as the evolution to your bot because people it’s, it’s starting from a disadvantage. Right. Because if you say it’s what you, you’re used to, but only better people are explicitly avoiding bots today, or even developing some of the words that were being used are in along the lines of loathing, right? Loathing is a very big word, very harsh word. So you don’t want to get from loathing to, Oh, okay. So this is the thing I loathe 2.0. I don’t think so. So understanding those perceptions is super important to developing your agent strategy.
Greg Kihlstrom: As you mentioned, most CMOs out there and most people in marketing probably have a good understanding of their brand persona or whatever, but you’re talking about something relatively new and pretty new for any brand, really. There’s a few out there that you mentioned that are doing things like this, but what does a marketer need to think about, like, is there something different they need to think about in their brand? Is there another element that they need to start thinking about as AI agents are becoming, you know, more and more the norm? Like, are there are there additional facets of their brand that they need to think through?
Gil Rosen: I think what we haven’t thought about in the past that this is because we just didn’t have the capability is the fact that we actually have now available to us what I call the option space of personalization that we never had before. So we can potentially put in front of a person the exact agent which would match the liking of a specific consumer, right? Based on, again, the history, the context, and everything. But is that what you want? You have to make a decision. Do you want there to be an infinite amount of variations to your agent, or do you say, this is my agent. You and I, we don’t know how Siri looks, right? So let’s imagine tomorrow Apple launches Siri with some type of facial or physical instantiation. So we know how Siri looks like. Do we want just one Siri or now is Siri evolving to fit exactly what I like visually and how I talk? So we have this, we call it the option space, the agentic option space. It’s, it’s, it’s infinite and we have to draw the circle around exactly what, how we want the interaction to take place. What is really important to understand is that from a service perspective, we have been used to being served in buckets, right? On segments. It could get, let’s call it creepy if. It looks like, you know, too much about me, but it’s actually not the case. So we are dancing on this very fine line of being super, super, super personalized on one side or being able to on the other side, keeping your brand persona and image in front of the customer. So this is kind of the balance of marketing in general and the execution thereof has to be, has to be good. But the first decision is how am I handling this? Am I always in my, I’ll give you another example. Do you want to allow your customers to customize the agent? How they look like, how they sound like, which controls do you give the customer? It’s another brand decision, right? I can tell you the, there are two parameters, which were really important for people. For the beginning for the, because we’re at the beginning of this phase and, and these things have to be looked at the year from now. And I’m sure there’ll change. From a look and feel perspective, there’s still this phenomenon that’s called the uncanny valley, which means that if the avatar, if you give it a physical liking, if it looks too much Like real life, people kind of feel freaked out about it because they know it’s an AI, but it looks like a human being. So they feel uneasy. So this is something we have to avoid. Avoiding the uncanny valley, I think will be temporary because Technology will get so good and people will get used to it. A year from now, you could talk to a person or an agent that looks like me and it’ll be okay for you because it’s, this is how it is, right? Tech evolves. The second part, which people really think that will make them talk to agents is if the agent will talk to, not only in their language, but their specific dialect and accent, which is actually kind of amazing because if you think about it, no matter where you come from and how you talk, Your brand experience will fit who you are and you won’t feel that you’re left out in any way because the brand talks the, let’s call it the official English or the official French or, and it’s not from your territory or, you know, neighborhood or whatever it is. And with agents, we have that capability. So it’s actually something that will bring the brand closer to the customer.
Greg Kihlstrom: Yeah, yeah. So the idea here is, you know, I know that customers are already familiar and not so thrilled with the bots, as you mentioned, for the research. But as agents take over more and more, this first point of contact with a brand is likely to be an AI agent, which means, you know, they’re taking on customer care, they’re taking on sales roles for the organization. What do you see as either the biggest opportunity or biggest challenge here for brands in creating the type of experience that they want to when it comes to these critical things, whether it’s saving a relationship, maybe there’s a rocky moment, or finding the right moment for an upsell or cross-sell opportunity or something. What should brands be most aware of? First of all, they need to know the customers are ready for it.
Gil Rosen: And again, we’re all regular people interacting with technology on a daily basis. And AI is there all the time outside the realm of let’s call it established brands, all these different apps. And we have a lot of experience to it. You know, I have conversations with my AI. I know it can carry those conversations. It feels pretty natural. Uh, we actually, we actually, within the research, we actually looked at different, uh, events or scenarios that, uh, from care, technical support sales. And customers have a very high expectations from the agentic experience in the sense that they believe that they can receive resolution faster and better than human beings. And this is based on experience. Think about it. The notion of waiting on the line for the next available agent will disappear. The next generation will not know the notion, the recording of, you are number five on the line. Your waiting time is, will just be extinct, right? So first of all, that’s a great improvement in customer experience to begin with. Now, you’re getting on the line with somebody that has the entire knowledge base that is available for the company. is not, uh, is not impacted by the fact that I had, uh, I woke up with a migraine or my girlfriend broke up with me, or I’m just, uh, not feeling well, or something happened on the subway on the way to work. So I’m grumpy and I have to now fake it for the call. I’m not going to get impatient with a customer that might be impatient with me because I don’t take it personally as a, as a, as an agent. So the level of service that we will get will know, like, uh, for sure be much better than getting it from people, from real people. And also as the fact that the AI is trained on thousands, if not hundreds of thousands of call history and understanding exactly the right time for maybe offering a promotion. First of all, being aware of an available promotion and knowing exactly when to turn a care call into a upsell call. Is some, I believe will be a skill that, uh, that, uh, agents will have and will be able to execute it in a way, which is at least on par with humans. But this is where it gets interesting. We talked to two different populations in the research. The first population was of course, consumers globally. And we also talked to, uh, to executives globally. And the expectations the executives have from the agents was from the, uh, from, from the AI agents was to be as, uh, at least as good as the people that are serving, you know, the customers today, where customers said, when you launch this, I expect it to be much better.
SPEAKER_02: Hmm.
Gil Rosen: Because as far as i’m concerned i’m not talking to something that knows it all right so my expectation is higher my expectation is higher from first of all availability twenty four seven. The speed at which I reach it, I don’t have to wait. The extent to which it gives me answers for everything I can ask for. And it’s not like you land in a care center and then they say, oh, let me just transfer you to the salespeople. Oh, sir, I understand you have this technical issue. Let me now transfer you to the technical expert. No more of that, right? And you have a complete record, of course, that you can actually work on the fly. So the escalation. route will actually be to human experts, but the bulk within the years, very few years, the bulk of the care and customer engagement scenarios will be dealt with directly with agents, AI agents.
Greg Kihlstrom: Yeah, I mean, so I think on alone, like, getting rid of hold time and wait time. I think that’s not only a customer experience, I think that’s an improvement for like society at large. Well, so I’m sure a lot of people listening are looking into things like this, or at the very least they will now after hearing more about it. What’s a step that a company can take to get ready for that? I would imagine there’s some things that are easier or make this adoption process easier of agentic AI and things like that? What’s something that somebody at a brand could do today to get prepared for this?
Gil Rosen: Well, the first thing I would urge the marketing people to storm into the, to the room where these decisions are being made and claim the territory that this is not just a technical improvement to the customer experience. This is a fundamental change in the customer experience. Therefore it has to be defined by marketing and brand aspect. So. You have to make a conscious decision of how are you going to manage your agentic experience? And, and by the way, you can evolve it. You can start to say, you know, we’ll start with having a single persona, let’s call it AKA Siri, AKA Alexa. Yeah. Just as a single persona, we’ll call it whatever. And it, he, she, whatever it is, we’ll be interfacing with people. And these are the controls we give it. You know, it’s a, it’s happy because we’re a happy brand. It’s a, it’s serious because we’re a serious brand and basically define the, you know, the, the, the characteristic of the brands and embed them in the, in the agentic experience. Then of course, depending on where you are from a. Territory perspective, or maybe you’re a global brand, you have to realize that you have to, you have the potential to serve your customers on a much more personalized level. So how, how deep do you want to take it? There are three circles you have to define. The brand is the one I just mentioned. The second one is the context. You serve a different type of personality based on the context. So a salesperson behaves differently than a care person. which behaves differently than the tech person. So whether it’s a tech support care support or sales opportunity, do you, you know, how do you fine tune the characteristics of the agent? And the third is the persona in front of you is he or she old is, you know, what’s the call history and, and, and basically customizing it. So you have to decide on each one of them, how deep you want to go. and basically decide your option space. And again, you can start simple and grow. What you have to avoid is A, and this is a recommendation, it’s not a fact, it’s a recommendation. Do not evolve your agents from or don’t communicate that this is an evolution from your bots. Disassociate from bots. They’re hated. Let’s not do that. The second thing is you have to know The two things I mentioned, one of them, which are important to customers, the language and dialect I mentioned, and the other one, a recommendation from a friend is very important for a good experience. So when you do your go to market, consider the fact that social circles are critical in the cascading or to cascade the message about your new experience. And, you know, marketeers can take that anywhere, but that’s kind of a overarching guideline.
Greg Kihlstrom: Yeah, yeah, love it. Well, Gil, thanks so much for joining today. One last question for you. I like to ask everybody before we wrap up here, what do you do to stay agile in your role and how do you find a way to do it consistently?
Gil Rosen: You know, I think, um, we live in a very special time, um, where technology impacts every aspect of society. And I think for me as a marketeer, who’s also, or always been in the domain of marketing and technology is that you have to remain amazingly curious. Everything has to be interesting to you. You have to investigate the impact of everything that is happening on everything that you do. And also look in retrospect and be able to accept things that you did, which were not the right thing. So, you know, I made a mistake a year ago when I did not even consider the fact that there is a link between this new technology and branding. I just did what looked good. I kind of instinctively did it, but didn’t like do it in the right way, which You know, so you recognize you made a mistake and you, and you fix it and you’re humble enough to understand we’re moving fast, but in this exponential era that we live in, we have to remain humble, curious, and always, uh, I’m, I’m, I’m on a personal level. I, I, I hate the status quo. So I am always looking for what’s next. And I think it’s, it’s very important in the age that we live in and it’s kind of, uh, how I try to, uh, to stay where I am.
