Lob's website experience is not optimized for Internet Explorer.
Please choose another browser.

Lobcast Podcast: Marketing AI & Appletinis

AI is a topic that’s on most marketers’ minds these days. So we sat down with Pini Yakuel from Optimove to discuss how marketers can embrace AI to create successful marketing campaigns.

Listen to podcast

On this episode of the Lobcast Podcast, we’re talking all things Artificial Intelligence and how it’ll change the way we do marketing. Joining us is Optimove’s Founder and CEO, Pini Yakuel, to discuss AI’s role in marketing.

Key highlights include:

  • What marketers need to consider when integrating AI into the marketing strategies and campaigns
  • How Optimove leverages AI technology in its platform 
  • How AI can personalize marketing campaigns and use customer data effectively
  • How marketers can find the right balance between human decision making and machine automation

Meet the Speakers

Stephanie Donelson

Senior Content Marketing Manager

Pini Yakuel

Founder & CEO

STEPHANIE: Hello. And welcome to the Lobcast Podcast: Mixers and Marketing. I'm Stephanie Donelson, your hostess with the Marketing Mostess, and I'm thrilled to be joined by Pini Yakuel, Founder and CEO of Optimove. Pini, do you mind introducing yourself to our listeners?

PINI: First of all, hello, Stephanie. Thank you for having me. And yes, I'm Pini Yakuel. I'm the founder and CEO of Optimove. Nice to be here. And I'm calling you today from Tel Aviv, Israel.

STEPHANIE: Well, thank you for joining us. And listeners, if you wanna make the complimentary cocktail for this episode, which is an Apple teeny, mine turned out pretty vibrant, but you're gonna need one and a half ounces of apple vodka, point seven five ounces of fresh lemon juice, and point seven five ounces of simple syrup. You're gonna combine the ingredients in a cocktail shaker that's been filled with ice, shake, and then straight into a chilled cocktail glass. So cheers and welcome to the show, Pini.

PINI: Thank you. I wish I had one of those myself.

STEPHANIE: Great way to end the day or start the day wherever we are in the world. Alright. Today, we're gonna be talking about a topic that's on many marketers' minds, AI. Some marketers are eager to embrace the technology. And some are kind of waiting to integrate AI into their marketing tech stack. So, Pini, Optimove has been here for more than a decade. When did you guys first start using AI in your platform and your solutions?

PINI: We started from the get go, essentially, Optimove was founded by a friend and myself. And at the time, one of my friends from academia, so we were both researching machine learning and AI, at university He had a PhD, and I was the dumb guy with the master's, with the master's degree. And, yeah, that's how we started. So we went out to the world. There's, like, two data geeks that are looking to basically harness the models that we've learned to real life problems. And we landed, we landed at retention marketing. So kind of like, helping brands maximize the value of their existing customers. By leveraging, you know, data and machine learning to understand the customers better and then to action that with the more personalized and relevant messages across all the channels and you know, that's who we are. We like to say that we are the the first, customer-led marketing platform. So that's that's Optimove in a nutshell. And today, it's four hundred and fifty people across New York London, Tel Aviv, and recently partnered with Summit Partners, which is our main investors from Boston and and, yeah, excited to be on this journey. And, you know, it's my first job. So out of Uni, doing this, and I can probably say that I was the one forging Optimove over the years and vice versa. I think the company helped me become who I am today.

STEPHANIE: That's excellent. And, I mean, I think as a marketer myself, we're always looking for ways to personalize our marketing messages because that's really what it's all about, making sure that you're delivering that right message at the right time to the right customer, and you need that data, to not only deliver that message, but make sure that it's personalized and relevant to the person. We can all spot a generic offer from ten miles away.

PINI: Exactly. Exactly. And I think, you know, as, you know, there's a lot of cultural references, right, minority reports with who has come to mind. And, but, you know, the the beginning, like, people never did. It was just too expensive and too difficult. To do it. Right? So I like to go back to Henry Ford with the T model. And like you said, you can paint it any color you want as long as it's black, So, like, I'm not dealing with that thing. Right? I'm giving I'm making all of my cars black. And this is, like, a civil, this one is not even paint. Right? And at time, you know, technology improved, then you got you know, the the the development of the assembly line and robotics and how you how they assemble cars into a, that you can order your there are Mercedes or Audi Q seven with, like, fifty different types of wheels or or rims or or you know, steering wheel or different patterns.

STEPHANIE: Whatever color you want it to be.

PINI: Well, you can even design your own Nike shoes, right, with your own, so and as it comes to marketing, we we empower our users to basically delight their customers with the most relevant message and personalized message to the customer.

STEPHANIE: So since you've been in this journey for a long time, how have you seen the role of AI in marketing evolve?

PINI: I think, obviously, like, around when when did the big blow up of OpenAI happen in ChatGPTs, like, six months ago?

STEPHANIE: Probably five months ago. Everyone's an expert.

PINI: Right. Definitely since then, you know, where you can see a step function. So I think, like, here, there's always, like, a very know, there's always been, like, small gradual improvements, but I think there's a few major step functions. Right? So I think, the first one is probably the democratization of computing. Right? So before before be because those things inquire a lot of, a lot of, you know, compute power. Right? So even if I I think the math was there, for a long time. Right? So, of course, the math is getting better as well. Right? So watch you. The math is is improving as well, but if you actually think about it, even at the before chatGPT and stuff like that in my company and in and in companies similar to mine, typically, the biggest challenge to actually run machine learning models was what you call ML ops. Right? So the operation of of data and computers that requires to ship out a predictive model into production and all the machine learning model into production. And at the beginning, only client only big, you know, the huge companies like Netflix and Google and Facebook, like, these are the in Amazon, of course, and Microsoft. Like, these are the type of behemoths that actually enjoyed a very strong computing infrastructure, what namely then after became AWS and GCP and Azure and all of those now they're actually setting it. So cloud computing is a big part that that was a step function. And now then companies needed to build a proper ML ops infrastructure to support that. But then chatGPT is indeed step function a hundred percent. And, yeah, I also appear when you, you know, bring me back. I I tend to go on tangent, so just bring me back.

STEPHANIE: No worries. I love listening to what you have to say about it. I mean, it's a very new topic. It's something that even I'll admit I've been hesitant about because At my core, I'm a content marketing manager. I'm a writer and seeing these machines coming in to take my job. I'm like, but but I like doing the writing.

PINI: Yes. That's, the yeah, for you. I'm actually not that great at writing, so I actually love it.

STEPHANIE: Don't worry. You're in good company.

PINI: Yeah. Yeah. So, you know, and I think I think look, with it's it's important to understand that, you know, mostly what we see, like, the the big step function that happened now, it's mostly about, large language models or how they call them LLMs. And and basically pictures. Right? So AI right now, so basically, it can understand language. You know, you can chat with it, and it's really smart. And, and, you know, it can write things on it. So, so kind of like writing has been decoded, right, cracked. So machines can write. Right? That's that's a big change. Yeah. But mostly it's it's around language. A lot of the a lot of the the improvement, it's our language, but also about, kind of, like, more than that automation and and things you can do and how you can take language and then perform a sequence of automated tasks because of language and, and, of course, and, of course, computers. But I think from our perspective, we usually use machine learning mostly to analyze data. And and that part is also has greatly improved, but I think it's fair to say that, this big step function is is more one on language and pictures.

STEPHANIE: Definitely. And I think, you know, I haven't been on any marketing team in my entire career where I've come on board and they're like, our data is perfect. It's one hundred percent clean. And so that's another big part of AI. It's if it's garbage in, it's garbage out. So you still have to have that human component of making sure that your data is clean. It works, and that way you're gonna get the outcomes that you expect. But going back to your own company, Optimove, how far are you planning on taking AI in your platform and solutions? Do you guys have a roadmap for that?

PINI: I think I think the question is not how far you take AI. Right? AI eventually, it's a tool. It it's now probably a more powerful tool in your arsenal. But ultimately the question is, can it help you to realize your vision? And we've always had, you know, obviously we wanted to talk. Right? So for example, for us, we are excited about whether I can have, like, this big search box within Optimove because we we host a lot of data, and we we transform this data into very friendly form for the market. So architectures don't need to know SQL or stuff like that to gain really deep insights from their data inside optimal, but they still need to push a few buttons and go into this dashboard and filter this or that and, you know, click this or that. Imagine if they can say, hey, you know, what's been my best performing segment in July 2021? And why? Right? So let's say so and then they just get an answer with a nice little chart. And So this idea of just being able to ask a ask a question and get me when you sit on top of such rich data.


PINI: And then just, you know, have any answer that you want to to get, you know, the speed from question to insights, I think, becomes much, much shorter. Right? Right. And so that's what I'm excited about. Some people are excited about innovation in UX. Right? So many people just saw, like, what Shopify did, right, and they're saying that they're gonna be releasing a new AI assistant in I tried it by. So so it's, like, having, you know, like, like a butler. Right? So I'm sorry. I'm a big Seinfeld buff. So, So when they use the board instead instead of a sidekick, but it's kind of like you have your own, and you say, hey, let's take the home page and change it from know, mostly focusing on jackets to focus on, you know, pants or something like that. And then so a lot of those tasks that you can do, but then you need to learn how to do them, you need to, during the platform, know which buttons you need to press within, specific UI and the term how to walk the UI. Now, basically, with a prompt, you can just say, do this and this and this, you know, and just chat with that AI, and then all of those tasks will be automatically performed. So in a way, like, if they well, the future may be system no longer have a UI with, like, a menu and, check boxes and forms, wizards, and you fill out things, they takes you to a place, you can just, you know, chat to the prompt, and those things will just happen. So maybe in the future, you know, all UIs just look like Google.

STEPHANIE: Yeah. That's a future I'm interested in as a marketer.

PINI: Yeah. This this is definitely because there's already, you know, HubSpot have been doing some things around that that you can basically just search and it would do some things for you I mean, I think there's still questions. Like, would you would the user prefer to because even when you prompt, you need to you need to provide all the details. Right?


PINI: So would you rather provide all the details in natural language? Would you rather just go over the because like, a form, for example, it tells you, it's okay. And it does anything you need to, you need to. But then the the check can come back and say, yeah, you missed those two. What are they? And then you can answer. So I don't know, but that's a that's an interesting thing that will happen, I think, around UX. I'm low I'm more interested in using it for data and for insights as I said before. This is my and, obviously, on the things that, like, people have done before, arounds, like, even in the space of marketing, you know, people that have been using AI for better copywriting. Right? So You had those use you had those use cases that you can, for example, you know, generate more alternatives from a subject line in an email. More alternative for a copy, like, is the main sentence in the banner.


PINI: So think about AI generating, like, you just give it the first one, and then it's generating, like, ten more alternatives.


PINI: And in the end's gonna be testing all of those ten and find the best one for your audience.

STEPHANIE: I love that. We actually have a blog post on our own website talking about should I use to write my direct mail marketing copy and we kind of hedge our bets and say like maybe do some you know like just copy tests. Like you're gonna run an A/B test split of this postcard campaign. Here's your header over here. Now generate me a couple different examples and then we're gonna move that over there. After giving it a review to make sure that the sentence actually reads correctly.

PINI: Yeah. No. A hundred percent. I don't think I I feel like if you think about things probably get better, but but I do agree with you that at the end of the day, I, you know, I think the way to look at AI is ultimately know, it's it's a very old cliche, but I still use it because I think it's very relevant. I think, like, from a broad perspective, you should think about the human. So when when I was, like, big sci fi things and, you know, Isaac Asimov had all of those books about the few and, which are now becoming very relevant. It's always about the machine don't know how to ask a question. The machine don't know how to design and experiment. Right? So those things are things that, like, humans still do better. And if you think about if you think about, like, you say, okay. The human the human user needs to define, like, the playground or the framework. Like, this is where things are happening. This is what you're now tasked to do. And within that, please optimize. So you own the machine, the the machine to optimize. Right? So, and it's because machines can run, you know, a lot of permutations and a lot of calculations that are very hard for humans to do. But humans can we can use our narrative driven brain to which this is something that we ought to that's what we do this home, you know, what people think about, you know, in in the in the shape of the narrative, and then we can design an experiment or ask a good question or so it's it's probably like, who can who can create better prompts to ChatGPT? Maybe that's the skill of the future. Right? Or things like that, it's what will determine it. And I think a fine example is, you know, they always talk about Terminator versus an Iron Man Right? So is it are the machines gonna be like Terminator or is it like Iron Man? Right? So, hopefully, you know, the good side of things is like Iron Man. So it's like a suit you put on, It just basically enhances your abilities and not something that's trying to kill you.

STEPHANIE: Let's hope not. I don't wanna go battle with some robots, but since we're already kinda talking about finding that perfect harmonious balance between AI and human in marketing, are there any things that marketers should be wary of when implementing AI, like on their own or with the providers that they're choosing in their MarTech stack?

PINI: I mean, warning, I mean, I'm not, I'm not, I'm not, your life is, like, what are the cautions? I'm, I'm a I think, you know, you need to be basically moderated. Right? So anything you do


Pini: Kind of, like, study it, be curious, research it, you know, take some baby steps, kinda like, see how it works, you know, go deeper and be open to it. I mean, we don't wanna be the taxi driver that thinks it, like, he knows better than waste. Yeah. I think that's, like, we don't wanna be that. Right? At the same time, we want to probably so we probably don't wanna go to say, like, oh, I'm gonna replace my entire department with AI within six months. That's probably too much to the other side. So start experimenting, you know, take your problems, make your use cases, take some courses, you know, I'm trying to, you know, I'm I'm using ChatGPT myself, being Israeli with with, you know, janky English, it's, it's definitely a benefit. So so, you know, but just be open and try it and and see what it I think always being open and and you know, being closed off and and fearful, I think that's a better approach every marketer. And because you gotta get you gotta be a part of it. Right? You don't wanna be left behind. So probably can enhance what you can do. And if you are building your skill set and you'll continue to be creative. It's definitely, you know, there's no problem and no risk for you.

STEPHANIE: Yeah. What about the flip side of that? Is there a minimum use of AI that marketers should be using right now to stay competitive in their industries?

PINI: Depends on, on, I, probably, I wonder what, like, you know, content factories are probably the ones that are. Very much, you know, losing it and trying it probably more probably junior level more. I would assume that, like, savvy content marketer that's been writing for many years for sure. Know, I'm gonna spend more time editing it than actually writing it myself. Right?


PINI: But, but, yeah, I think everybody should be open for it because indeed. There is a step function.


PINI: You can't, ignore that. Like, there is a step function that I have. Right? You know, there's millions of people tried it, and they all posted on LinkedIn, like, see how amazing. You know, it's all this and this for me, and it it, like, people are impressed. So--


PINI: And it's always getting better. Right? So it's getting better and better by by by the week, I think. So I think definitely you're being open to it. Exploring it, experimenting, and, and see where it takes you.

STEPHANIE: I love that. Alright. We already kinda talked about personalization earlier because we know as marketers that personalization is so important. In fact, in our recent 2023 State of Direct Mail Consumer Insights report, it shows that personalization continues to be a key element for marketing and direct mail. Our report shows that sixty eight percent of consumers are more likely to engage with a message or communication from brands that are personalized to them. We all know personalization has come a long way in marketing. But, Pini, how does AI play a role? Do you think AI can truly deliver real-time personalization in our in campaigns?

PIN: Yes. A hundred percent. I think a lot of it is a lot of the happening. So, obviously, AI is it helps us to know our customers better. Number one. So by leveraging all the data that we can collect and gather on our customers. Right? First party data, So everything we know is a brand, zero modeling data, if our customers give us this data, right, just fill out forms, and if we can get third party or data from somebody else, at the end of the day, AI is so this is what we do at ops. We analyze the data and we put it in a in a specific structure that, you know, discovers interesting personas as a customer. Right? So if I'm looking at this this huge database, and I'm trying to make sense of it, don't wanna be asking a question, and how many people did this than this, and how many people bought more than this, and how many people they like this product. I want the AI to tell me, hey, this is an archetype of customers that you have in your database. Right? So you have, and it's, and it's coming out, like, something that may be in the past. You you needed to hire a McKinsey consultant to run a project of over eighteen months. And do focus groups and panels and things like that. Maybe you can get those archetypes with AI much more faster, of course, panel makes it an event because, you get different type of data there. But so this is something that's already happening speaking of, you know, giving it in real time. So, exactly, the example we used before, right? So if if I'm in, Another thing that we use AI to do. So if I'm basically testing out a few alternatives for marketing campaign, I can have an A, C, and C. And usually, how people used to do it. You run A, B, and A, B, and C, and you look for the winner. So you say the winner of this campaign is is permutation a. So now let's deploy permutation a for the entire segment.And what we do instead and have been doing for a while, we have this thing called a self-optimized campaign where we are basically deploying A, B, and B, but then we study the winner per microsegments, not the linear thing. So that even the losing even the losing action, let's say you have one of the variants of the campaign is by far, the walls performing when you test it on the entire big segment. But but you have four microsegments that love it. So that's what they should see.


PINI: Right. That's that's personalization. So again, you we use this approach called contextual band, contextual bandit to, essentially solve this problem, and we leverage all the deep data that we have in Optimove and microsegments and things like that to be able to to do that, and to close the loop with all of the machine learning ops and and the things you need to do. So ultimately, And those things are already happening. And I think in the future, some other things that we're looking at, like, whether the style of writing will impact your propensity to engage. For example, I don't know, Stephanie, if you let's say you have a brand that you really like to buy from.


PINI: Of course, that brand has a specific voice. Right? So some brands have a specific voice. But then they have a few permutations to that voice? Right? Is it legit that you will get a message that's basically more clean and formal? And I'm gonna get a message that's maybe more whimsical and humoristic? Now the land the the the overarching message of that campaign, read this thing. Right? They're gonna tell you and me that there's a new product line, and we both get a discount if we buy it, by Friday.


PINI: But y'all is gonna y'all's is gonna be more formal and mine's gonna be more, or maybe mine's gonna be New York style and y'all's gonna be Boston style. In the copy. Like, would that work? I don't know. Maybe some research shows that it does.

STEPHANIE: Oh, yeah.

PINI: So now it's very that's very easy to do now.

STEPHANIE: No. And I've done that before, you know, again, being in content marketing, You really have to understand how the style, your voice, your tone. It kind of can change the message a little bit and how the person perceives it and what they take action on, especially on a per campaign basis. Since we just kinda talked about personalization in real-time, let's say through a specific campaign, what about bigger marketing strategies? Do you think that AI is able to make strategic marketing decisions.

PINI: I do not.


PINI: I do not. No. I mean, I think strategic decisions require semantic knowledge of the world and your specific business and the old specific goals, which I think it's not not yet. Right? Maybe that's maybe the maybe the next step function, right? Maybe ten years, fifteen, twenty, for thirty. I don't know. But, no, this strategic decision of course, it can help inform the decision, right, and, and guide the decision, but the decision and the strategy will be done by, you know, a set of molecules called a human?

STEPHANIE: No. And I think that's a very fair answer. Right? You A machine is just making decisions based again on the datasets that you've given it. So it understands, okay, if this then that and moving through, whereas a human, yes, we are still doing that essentially, but we're able to bring that historical context. We're able to see the bigger picture. We know things that the computer doesn't. And I agree. I think that we should definitely be in charge of it coming up with a strategy, but then when it comes down to kind of those campaigns, allowing the AI to do what it do, what it does best, and using that data to make those real time personalizations, to make that change, to saying oh, hey, you know, John, John Smith in New York needs to get this version of the campaign instead because we've seen him take actions on ones like it before.

PINI: Exactly.

STEPHANIE: Alright. So, Pini, what advice would you give to marketers who are looking to leverage AI in their marketing efforts? But are kind of worried about relinquishing control.

PINI: Yeah. That's Yeah. I think that the control pieces is actually pretty big. It's funny. Like, we find that, maybe I'll answer with with a simple, like with with an example, we're storing from our users with the user base. So let's say if we have a user base, that's been we have a client that's been that client has been using Optimove for four years. Right?


PINI: And they they have a few habits. Right? So they're used to doing something So for example, we had a process of decisioning, and that process is manual. So the user needs to choose in the GUI, they need to say, I want this to happen before that, stuff like that. Now I'm gonna release a new AI feature that takes this manual process and makes it automatic. Right? So AI will make the decision forward. So the users that are used to the system and use and have been using the manual process for, like, two, three years, they will not adopt a new AI feature. Okay? But a client that just owned one of the month ago, They don't even know that we had the feature of doing it manually. So for them, when they get started, for them, okay. Fine. So this is one more thing I just don't need to worry about. So because it's automatically provided to them that that decisions are taken by the machine, they adopted much more.


PINI: So I think it's it's similar to, like, to, like, when you see, you know, in younger generation versus an older generation, adopting some kind of a new technology. And, like, I just, you know, my uncle who lives in New York came to Israel, and I was like "Why aren't you using like Google Pay or or Apple Pay?" He is like, I don't know, you know, the little --

STEPHANIE: That's not how it's done.

PINI: I'm gonna so, like, I'm gonna, you know, get into my phone and steal my identity. I was like, I was like, look, it's actually safer because, like, a credit card is just a piece of plastic just to go in here. I need, you need your fingerprint, and it's not as as easy to It's like, but, you know, it's like that, but, but, like, a twenty year old is not gonna even think for one split of a second about they probably don't want the plastic at all. Like, they don't have it. It's somewhere in their home or something. So I think it's a little bit like that. I think at the end of it, but same degree. Right? You you need to I think to gain trust in anything. It's not necessarily a question of AI. It's a question of trust. So in order to trust something, I think you need to sample it. That's that's my so I come from, like, you know, an industrial engineering background. There's a field in methods engineering called work sampling. So in order to get a picture of reality, if you sample something, then it gives you and after after a certain amount of samples, love statistical credibility of a certain view of the world, right? So I think everything that I do and everything I see people are not doing executive was like, when somebody in order to master something, you need to get into the details. You need to you need to dig deep. Right? You need to open the hood, look at the engine, look at the wiring, you know, touch it, make sure it's not loose. And once you do that, you gain the trust, right? So if you haven't processed that's that's running with AI. Right? Experience it for yourself. Right? So let's say have new mock users, what are real users, have them experience in, like, they'll phone the company, let them see what campaigns they get. Yeah. Let's see. Let's see how it looks like on the other end. Right? Let's experience it. Go on to the database and make a few queries and see what actually is happening in, in reality. When you do that, you can gain the trust, right? You can understand what it does. Then if you want to get continuous trust, build a few monitors. So get monitored that, like, sends you a message of something go above this value. Right?


PINI: So with with with with basically looking in looking deep in having some monitors, getting acquainted with the data, do some work sampling, do like this mystery shopper type of a thing for your own service.


PINI: That's gonna cover you.

STEPHANIE: I mean, there's a reason so many technology companies offer those free trials, the personalized demos where they ask you for some of your data so they can actually show you how it really works. In your own instance.

PINI: Right. Right. And, and I think and but but I do think yes. It is important too. Many things that you're gonna buy are probably gonna be buggy, and they may not work, and they may not do what they say they do. So just, yeah, just just kinda like experience it.


PINI: Test it. Don't be afraid to look deep into the data, see how it works from you, and then you can have the trust.

STEPHANIE: Definitely. Alright. Pini, do you have any real life examples or success stories that highlight the impact of Optimove's AI powered marketing platform on your client's customer engagement and business growth?

PINI: Yeah. For sure. So, what we see kind of like what when our clients are really successful with us. Right? And and when that magic happens, it should be a combination of a few things, but when it happens, what we see is number one, we see first of all, we see a consular transformation to review the marketing team becoming much more data driven. We see them becoming kind of like, in case of like sedimentation, and we see them being completely free to have new ideas and implement them in, in minutes or hours versus weeks or months. And and and ultimately, what's happening is when you start to delight your customers at scale and where you start to provide this personalized experience at scale. It only happens when you do it at scale. When you have a one-off person in there it doesn't really work. Due to have kind of like, we call it from tens to hundreds of segments. And you can manage that at scale, and and, and that's what optimal does. It enables you to do that. When that happens, you can see that. The way we like to measure it is is we we we call it like what's the CRM marketing contribution. So if there's a business that's doing a hundred million a year, and we can we can execute ten million of that hundred million to theory marketing We call it ten percent CRMarketing contribution. So and the way we can do that is because everything we measure, we measure based on incremental So I can come in and say to a brand. Sierra marketing this year did thirty million dollars. And then they can see what portion is that of their overall business. Now that's not only optimal. Right? That's optimal. It's also this CRM team and the design team and the merchandising team that created the promos and like this, it's a but we just were able to measure the impact, and, of course, you guys were delivering the the great, the direct mails, which which can be one of the most impactful channels. Right? So so it's leveraging the combination of the channels and the intelligence and all of those things together. We can measure that uplift. We can measure that incremental value.


PINI: And when it's really good, it could be up to thirty percent. Of of a business of a business as well.

STEPHANIE: Oh, I love it.

PINI: With CRM Marketing, it said it's best. It could be thirty percent that you can measure. And, of course, don't forget about what it does to the brand, which is unmeasurable. So when customers are so happy when they're getting those personalized experiences.

STEPHANIE: Well, that just goes back to customer retention. If you're keeping customers happy, they're going to stay with you longer.

PINI: And gonna tell the friends and then it becomes an acquisition channel.

STEPHANIE: Yep. Oh, we have talked about that plenty of how to use customer retention campaigns as kind of a dual campaign in customer acquisition, the, refer a friend, the friend friends and family discounts.

PINI: So so you'll get like a so that's really great.

STEPHANIE: Alright. So we've kind of talked about the challenges around AI. But let's talk about the benefits. Like, what are some considerations that marketers should keep in mind when incorporating AI into their marketing strategies? Or what benefits should they really be looking for to pay off with that AI?

PINI: Again, I think I think we covered most of those things along the way, you know, with different conversation topics. But I think, again, in general, it's like getting more insights from your data getting those insights in an easier fashion, in a more digestible, palatable fashion. Helping you with copy, you know, finding winning copy, finding winning images, and, you know, everything is is, you know, that this whole process of test and learn.


PINI: Which if you can do it, you know, if you can do it at scale, again, it creates it just becomes a different culture. And I think today it's very clear that, like, test and learn, is, is really is really a big part of winning companies. So for me, kind of like being Israeli, there's very. I can I can share something cool with you? So, in Israel, in the army, then we always say kind of like the best will become the air force pilots. Right? So the air force pilots are like, and it's like a really strong brand, right? I'm I'm probably in the US as well, Top Gun, stuff like that. Right? But for us, like, the top gun is really, it's really big. Right? And it's very famous their culture of the of like incident reports and, and basically you know, going back on everything they do and learning from their mistakes. And of course, it sounds very natural today. You know, people know about it. Most mistakes, if something bad happens, I'm gonna ask for an incident report, but every air force pilot that I talk to they always talk about how this was monumental in shaping them and shaping their style as leaders, as managers, as people is this notion of movement, constant learning, this this this idea of of test and learn, right, always always always a huge part of who they are, and they brought it to the the Israeli business culture. And I'm sure when I always when I engage with American companies, you you can very vividly see that this is a big part of the culture that has to learn AI.


PINI: Supercharges that. Right? That's what it does. It supercharges that as well.

STEPHANIE: Yeah. And I think that just brings us back really nicely. You know, we've been talking about using AI to reach out to your customers. We're all about telling stories and I mean stories at the end of the day are all about transformation. And change. And that testing and learning leans right into that, like, you are gonna get left behind if your company does not adapt and take advantage of these new technologies to just improve what humans have already done before. So we've already had the creative minds think of these campaigns. Okay. Now, how can we optimize them without spending our time manually analyzing that data?

PINI: Exactly. Because it's because in many cases, we don't know, right, the reason we need optimization and testing is because, you know, even the smartest people get it wrong in many times. That's the bottom line. Right? I can, you know, I can tell you this, like, when I was younger, you know, being kind of like, you know, a young cocky individual. They're just kind of like left university. I was, you know, I was, I was certain of a lot of things, and, you know, my experience taught me that, like, sometimes I had a really strong conviction about something turned out to be completely wrong because, you know, reality is stronger than us. Right? So The proxy mentality is just being open enough to, to have those ideas. Let's see which one sticks. Right?


PINI: But that's the box that works well.

STEPHANIE: Perfect. Alright, Pini. I have to ask, do you have any final thoughts that you wanna share or is there anything that we didn't get to today as a concerns AI in marketing?

PINI: Maybe you can you can basically edit out the fact I said that I was a young cocky individual.

STEPHANIE: I'll see what I can do!

PINI: That would be my final thought there. No. I'm just kidding. No. I think I think we're able to cover a lot and and, you know, talk about this very interesting and riveting space. So thank you so much for having me. And, let's see how it plays out.

STEPHANIE: Alright. Well, thank you so much for joining us. And to our listeners, Thank you for joining us for mixers and marketing. Be sure to learn about more be sure to learn more about Optimove at optimove.com. And if you wanna dive deeper into the topic of marketing automation and AI, particularly in the direct mail marketing space, please visit lobdemo.co/directmailai. That's lobdemo.co/directmailai, direct mail AI is one word. As always, you can browse our library of episodes over at lobdemo.co/lobcast. Otherwise, thanks for listening, and that's all folks.