Scale AI Adoption: A Playbook for Approvals

Join  Adobe Experience Maker Award winner Anish Raul  for a candid conversation about bringing AI into large, highly regulated enterprises without slowing down or compromising compliance. Instead of just showcasing results, Anish reveals his proven playbook for moving AI ideas from concept to enterprise-wide approval. Drawing on years of leading personalization and AI initiatives, he’ll break down the repeatable process: how to prioritize high-impact use cases, build trust with legal and security, and embed principles for safe experimentation at scale.Anish will provide his own use cases with Adobe Target and the Adobe Experience Platform AI Assistant as examples; the strategies and lessons will be relevant for any organization wrestling with legal alignment on AI.

Transcript

learn from your peers webinar on the topic of scaling AI adoption, a playbook for approvals.

I’m Susan AI Aziz. Just kidding, it’s actually Aziz, but I had to make the AI joke since that’s what we’re here to talk about today. I’m very excited to be your host today. I lead the customer programming for Adobe Experience Platform, including our champion program, and this is a topic I’m super excited about and very passionate about. As marketers, our roles are evolving quickly and the content we’re sharing today is designed to give you practical, real-world tips, not just on securing approvals, but on how to keep building that momentum afterward through upskilling and driving adoption across your organization. You’ll see today’s agenda up on the screen here. We have a fantastic speaker who I’ll introduce in just a moment, but before we dive in, just a couple of quick housekeeping notes, and please don’t forget to respond to our poll. It helps us understand where you are in your AI adoption journey. While I walk you through a few housekeeping, I’d love for everyone to start getting active in the chat. We’ve got an awesome moderator ready to jump in with questions, so don’t be shy. Share your thoughts and questions as we go. To kick things off, let’s do a quick round of intros in the chat. Where are you joining us from? What are you hoping to take away from today’s session? And since we’re talking about AI, if your AI could take over one chore at home, what would it be? I’ll go first. I’m joining from New York City. I’m a foodie at heart, and if I could offload one chore to AI, it would definitely be folding laundry. I avoid it at all costs. Can’t wait to see your answers roll in. In the meantime, we’ll go over some housekeeping slides. All right, while you’re sharing in the chat, just a couple of quick notes here. So we are recording the session, including the AI chore debate, and we’ll be posting the recording and slides in the community afterwards so you can revisit anything you missed. Also, don’t forget to make the most of the chat. It’s one of the best parts of being here live. Feel free to ask questions, share your own insights, and connect with others. And finally, you’ll see a quick few polls pop up during the session. Again, we’d really love your input. It helps us shape the conversation and better support you. In fact, I’m going to launch a new poll here for the audience.

Awesome. And one last quick note, you’ll find a resources box with links relevant to content and more AI webinars that Adobe’s hosting. Our amazing speaker has provided his podcast and some very cool AI videos that he has created. So feel free to explore those as we go on or save them for after the session. I’ll take a quick look at the chat to see what I’ve missed. We got some folks from India, from Spain. And if AI could do anything, move my laundry from washer to dryer. I love that. That’s hilarious. Awesome. Okay, well, drum roll, everyone. Without further ado, I would love to introduce you to our speaker, Anish Rawal. Anish Rawal.

Susan, thank you for a fantastic introduction. Good morning, everyone. I’m Johnny from Denver, Colorado. We have an awesome room here today. And I’m so excited to be speaking with all of you. My name is Anish Rawal. I’m a senior marketing lead for personalization for Sling TV, which is one of the largest live TV streaming companies in America. But I did not always start as a marketing technology guy, as you will know in your story. I originally started working as a creative director and as a good digital marketing person. Done a bunch of cool stuff in the social media strategy, digital strategy, even producing like movie trailers and stuff for Netflix and some of the other streaming platforms I talk about. Lastly, my background has been at the game and focus.

My Adobe focus is the Adobe Experience platform because personalization target for the same reason for A B testing personalization. Adobe analytics, you got to tell stories with data to move the needle, guys. I was also named the Adobe Experience platform champion for this year. So it’s an absolute pleasure to be recognized and to be in this forum with all of you all. And as also an Adobe Experience Maker Award winner earlier this year for the disruptor category for some of the cool things that we’ve been doing at the company with our entire team. And I’m happy to be sharing some stories today. Some fun facts about me. So three pictures that give you a sense of who I am, outside of work. The first one on the left, that’s me catching my first ever trout in Colorado. Colorado made a fisherman out of me. Never thought I would be enjoying the outdoors as much as I do living here now because I was born and raised in Mumbai. Moved to New York City, lived there for a while, worked there for a while. The picture in the middle is from my time in New York City. That’s me with Casey Nicekak, a kick ass YouTuber influencer, a guy who really brought filmmaking to YouTube and revolutionized the daily vlogging platform. I started making vlogs and films because of him. Check me out on YouTube, just look me up on eShaol. And the image on the right, that’s behind the scenes for which we’re from one of our latest podcast episodes. Susan mentioned, I also have a podcast. It’s called the Art of Commerce, where we have conversations on creativity and conversion. It’s a really cool marketing storytelling podcast. So check it out.

If you should have the link in the resources. All right, let’s keep going. I see someone from real quick is a snapshot of my journey. I started as a copywriter in 2014 for MTV. Now, 2014 is the time when it was a time of click break. So I would be cast with writing things like, hey, seven reasons you should engage with Gemini or five reasons why Leeuw Maguri, Kipriowski, Hartrop, everyone should be talking about. All of those are cliche, these really cheesy articles. It was a good time because a guy was what the medium wanted. That is what people really like to read back then. So we would casually clock like a million hits in a week back then. And that was really my first foray into storytelling for the digital platform. Slowly evolved into a brand strategist. I was really lucky to work at WPP, which is one of the big four advertising holding companies. I was leading brand strategy for Netflix. So I’d just come to India and then my evolution happened from being like a writer to a social media get you good strategies. So think about Narcos, Stranger Things, all your favorite TV shows. How do you localize like an international video series to a local market and create content around that? It was around that time that I also made a big foray into video production. I used to make trailers for all of these movies and shows. Used to design posters with RDE man, evolved from being a writer to now like a multimedia production kind of gig. From there on, I went to lead the creative studio at Eros International, which was back then the biggest film studio in India.

And if you notice a theme in the first three gigs that I spoke about on this timeline, all of them are largely creative. And for creative people, I’m sure we have a few on the call or a bunch on the call. We largely leave with instinct, right? Does this feel good? Do you think based on your creative muscle and your gut will this perform well? And usually it does, right? The inflection point really happened for me when I joined my current job, which is a fortune 200 screaming company as a senior e-commerce guy. And I’m like, Hey guys, I have been a creative professional all my past decade that I have been working. So why do you need me to come in and even work on e-commerce, which is something I, which was fairly a new foray for me. And when I took up this gig, I realized, Hey, instinct alone doesn’t credit anymore. And I’ll tell you why, as we go through the slides. So the challenge character we were facing back then is, and even to a large extent, majority of screaming customers are not loyal, but they’re looking for the best value. All of us on this call, I’m sure we are subscribed to at least two streaming services. And what do we do when, especially if it’s video, most of the times we subscribe to something for some time, we watch what we want and we are out the door. Very rarely do we have services that we keep an ongoing subscription for. And when we are in that mindset as a customer first, that means we are also window shopping on multiple websites or, you know, multiple apps before we locking an option. And when the website is your number one sales engine, brand equity really matters. The reason I’m setting all of this up is I really need you to understand the context in which I came into this, this particular role. So lack of loyalty in the screaming space, customers are checking out all the possible options before deciding to locking something. And when that’s the scenario you’re working with, you really want to make sure whatever interaction you have with the customer on your website, sticks.

Now this is the scale of the site. When I first joined beautiful site for back then, right? We had on-brand images, but largely flat because, you know, we are a TV screaming company. TV does not look like this, right? TV has live dynamic images of real people, but when you are merchandising or, you know, putting imagery up on site, you have to navigate IP challenges, intellectual property, which is not the easiest to get because you have to navigate, permissions. You have to pay for it. And also like most companies, especially back then, our creative pipelines were backed up, which means to get a change of creative assets, you have to wait at least four weeks, which is fine. But you know, being from that creative background, it helped me understand why it was fine because really what happens when you request a new creator or a new image, you, as a designer, you go to a stock website, you look for the right image, you download it. And once they download it, you work on it, you do graphic design to bring it on brand, right? You make sure your color palettes, your tone, all of that is well suited.

And then you make a bajillion variations of it. One of which comes to my channel, which was the website, right? Now that’s the background that, you know, we were really working with, and this is the state of the site, but really the tipping point was shortly before football season, we come into office and realize, Hey, one of our direct competitors has the same image that we have on our landing page. Now, remember the problem that I was telling you about customers and not loyal, you’re probably going to check out this website and the next one and the other one. And if we are using the same image, we lose brand equity, right? And this was a very common problem back then. Why do I say back then? Because things are quite different now, and we learn Pat-Tac, but every brand and their competition was pulling from the same stock image libraries, right? And commissioning a custom for issue is a costly affair. It takes time, right? So these were the realities of the business that everyone was facing and say, Hey, what can we do about this? Cause you know, this is obviously something that is not the best. And we want to start to give the best experience to our customers. This is early 2023. This was also the time when Google trends was popping up. You could see it on your social media, Chang AI image creation. That was the big thing. And mid journey was really a tool that was popping at that time. And it prompted an insightful discovery. I said, Hey, can this help us solve this problem? Right? And as we went down the rabbit hole, I realized, yes, Chang AI image creation can help us solve this problem with speed, with quality, with price and speed, quality, price. Yeah. And, and really like all the challenges that I mentioned before, like, Hey, creative pipelines, my backup, the imagery is not unique to us. All of those could be bypassed. Right? So we started experimenting, but very quickly I realized if you want to go fast, go alone. If you want to go far, go together. Why do I say that? Because it’s very easy when you’re adopting Gen AI technologies to be the first going to be the only one who does it. But if you really wanted to scale across the enterprise, you need to work with people. You need to work with your stakeholders to make sure that you’re not a bottleneck when it comes to adopting that technology. And also the true impact of everything that I spoke about can be realized when the whole organization is adopting it. Right? So in the pursuit of going far, going together, we had a series of levels of second conversations or alignment conversations, largely because it was Gen AI creative production and our first brush with generative AI technologies. This is still early 2020. So about two years ago, right? We spoke largely with four teams. First one is legal. This technology is so new. And even today, we are trying to still wrap, largely we are trying to wrap our heads around how this is, what is the legality of this. Right? So in our conversations, we realized that what does legal care about? They care that each data set or each output is claims and right clears, they’re privacy compliant and backed by clear accountability if something goes wrong. I said, that’s a great concern and hurt. How do we navigate that? What does brand care about? Brand cares that it must protect our brand equity. It must unmistakably feel like us because brand teams spend years building that relationship, nurturing that tone, that unique identity, and you don’t want to damage the reputation of the task. Then there’s the creative studio who really takes the overarching strategy from brand and brings that idea to life by producing these creators. Creative studio is concerned with, hey, speed is great, but it must meet our quality. It has to feel like it was done by a human and it has to feel unique at the same time. One word or one term that is very common these days that fascinates me is AI slop. AI slop is just creative or like assets stack or really any output that feels half big and doesn’t really have that finesse which a lot of us professionals play on. Lastly product, how can we protect and enhance the customer experience? The website, the TV apps, the mobile platforms, if you’re planning to use these assets on all of those touch points, which by the way, millions of customers interact with every year, how do we make sure the customer experience is on an end-to-end basis? That really prompted a discovery and a lot of conversations. It wasn’t just one, but just a series of alignment conversations. How did we approach this? It was simple. We worked with each of these teams, we set them up with the technology, and just let them play with it. I still remember one of my first few weeks working with League of Legends, the guy was really the first step. The output was fantastic. I’ll show you a few examples as we go through. The output was fantastic, but how does this work? The last thing you want to do is introduce a black box, especially at an enterprise scale, which when something goes wrong, there’s no accountability for. I still remember sitting in my legal team’s office, setting everyone up with an account, and everyone is just creating assets. That was a fantastic day because for someone like me who loves creating or producing assets, it was just a bunch of people, not necessarily with their creative professional background, but all of us are creative at heart. Everyone loves just exploring social media feeds and trying to get a sense of what this can do. That was a great experience. Same thing with brand, with creative studio, with product. We had a series of workshops just to get everyone comfortable with the technology. What were the first conversations around Gen AI? Is it going to take my job? It was very clear to us from the start that this is not going to take anyone’s job. It will just make our job so much easier because now look at the challenge that we just can potentially navigate. You’ve got backup creative pipelines, stock images are hard to take a lot of time to produce and bring on brand. With this technology, we can bypass that. We can create high quality assets. We can produce it at a really great cost and just make it really unique so the brand equity and customer experiences are protected. A series of those conversations, not just with the people who would be helping us with the approvals and the adoption of using that technology, but also the leaders. We try to figure out what do each of these executive leaders want. Well, the CMO wants AI to be bold, but always on brand and growth driven. That’s a lower excitement right there. On brand, sure. We want to make sure it is consistent with the style, with the usual looking field that the brand usually goes for because it is curated and growth driven. That took us a while to figure out. We said, hey, growth driven, what do you mean by that? Really what that means is can this help us save money? Can this help us save time on a larger scale so that we are more efficient as an organization? Really valid concerns. The answer was yes, we can, but let’s figure out how do we do this. India is the chief legal and compliance officer. Like I said before, can we protect the company? Do we have the right risk coverage? How do we navigate the whole rights issue? For the chief product officer, hey, we need something that’s reliable, scalable and trusted by users. All of this happened with a series of conversations, with a series of just figuring out what the gargle is, what processes do we follow to scale this safely so that it’s not the wild west. We’re not running wild with new technology. How do we figure out access, governance, all the good stuff? And in a matter of a few months, we were able to do some really cool things. And I’m going to show you a few examples. We’re like, okay, what can we do with this technology? And we realized, hey, we can do hyper-realistic humans. So these really are the first few vendors that we ran with on some of our touch points. And very quickly we realized, hey, the output is something that we can, we are something we can be proud of. So our landing pages immediately in a matter of a few weeks went from looking like what we were looking at before to something like this more dynamic.

Skill on brand, very much premium, high quality and unique. You won’t find these images in stock libraries. And that was a really cool challenge to solve. And then we went further as we started grabbing adoption. We let our creative teams, our product teams run wild with this. And then really this one seed of an idea of, hey, can we use JENEI image integration for solving website photographs, evolved into something larger very quick, because now we had the right understanding, the right guard rails and the right, really the right people, the people processing technology, that’s something we talk about all the time. We had the technology which could help us solve for this, but really now we were trying to figure out the people and processes to really accelerate this.

And when you do that, you get greater option at scale. So this is a snapshot from one of the campaigns that this was like a year-end recap. We said, hey, can we create cute characters based on what customers like to watch? And then we were like, okay, yeah, cool. We can do that. So there’s someone who loves a baseball, a news junkie, like sorry, a news monkey, like a news junkie. We have a tennis puma. We created these fun characters and think about it. This is really common now, but back then to do this same kind of work a year ago would have cost you hundreds and thousands of colors, because you need to commission someone who’s an illustrator, someone who’s a creative rendered artist, then a graphic designer to really get each of these characters in. Now it was a matter of minutes. We also used it for UI inspiration on the websites. Earlier we used it in very disconnected pieces on the web, like, hey, let’s get a marquee or a hero from here. Let’s get a banner from another place. And then what we would get is a functional website, something that sold really well, but aesthetically had some ways to go. So can we use it for UI inspiration? And really what you see on the screen right now are just some of the first output, right? Some of the first few outputs, but then we were able to translate them into Adobe Experience Manager and really create digital experiences that were coherent. It had a certain theme, like on our screen right now, we see the one on the left, we call this concept like Basket Blue. On the right, we call it Firing Eyes. And just really exploring what is possible with this technology. Now, like I said, we were able to do all of this, but the backbone of this adoption really was some legal parameters that we were able to set early on with our legal experts. So here are some do’s and don’ts that we started with. We said, hey, you really want to create your own prompts by thinking about the output first. So do not see Wes Anderson in the style of Wes Anderson, because that doesn’t protect us legally. We want the prompt to be clean. Same thing for, you know, if you want to create a sports act, don’t say Nike because the sky is still low. That will get you the output you want, but you really want to steer clear of intellectual property in any prompts.

Come up with original copy to describe the idea that you wish to create.

Do not use any copyrighted messaging or trademark phrases. So those were largely like, hey, think output first, steer clear of any IP and always make sure you reverse search your imagery. That was one of the big ones. So we would reverse search our imagery, send it through the legal with the output and what comes up when you reverse search things. And that really helped us just identify, hey, risk is good for release. And here are some mounting settings, hexagrammatic pink Pascal mint. You know, we are really going for that Wes Anderson flatly Pascal colors, style, but we had to decode it and really make sure the output was safe. Same thing, avoid copyright traps by just, you know, yes, you may want LeBron James on your, on your header, but you cannot have someone that looks like him or you will, you will be a rebel for likeness. And yeah, that really, you know, those kinds of legal gargles helped us set those expectations and operationalize this technology. But very quickly we realized good images or good looking images mean nothing without IDEA. That’s a fun acronym. Thank you, Jack. Really that captures the essence of what we’re going for is those creative choices that you’re making. Do they have impact? What measurable impact did the creator bring? Is it grounded in data insights? Remember these CMOs ask of being growth efficient, right? Being, being efficient in driving growth. What, what can you justify with data efficiency? Does it bring about operational or cost efficiencies? And lastly alignment. Is it aligned with the larger brand vision? So really these image generation, this image generation technology was our first brush with AI. And we had to start proving these things out. So we get a bunch of testing. I’m going to quote just one that we did, but you know, we in very specific contexts and very specific use cases, the Gen AI imagery held up sometimes even outperformed, right? Like this example that you see on your screen, we had a TV presenter on a, on a basketball landing page. Someone who’s not as, not an athlete themselves, but it’s very strongly associated with the basketball sport. And on the left, you have someone who’s a generic AI creative. Now in the context of that landing page, in the context of, we are promising basketball on TV. For some reason, the celebrity imagery or the TV presenter imagery did not hold up as well as the creative, which was the edge of the sports player. So we were just optimizing and we figured out, Hey, this is, this helps us. The cost saving alone is super efficient and they also hold up when it comes to performance. And that was our first brush with the Gen AI as an org. My first brush with Gen AI as an org and very quickly we realized, Hey, what are some key takeaways when we want to try and adopt this technology at a greater scale? The first one is you have to go unified risk and compliance check, because if we are thinking of adopting these technologies at an enterprise level, you want to make sure we are aligned with legal and product. Make sure that copyright regulatory compliance, training data risks and usage guidelines are absolutely watertight.

The product has to validate that the customer experience is really watertight. We need brand safe, consistent creativity. So while brand protects core identity, tone and cultural sensitivities, we have to make sure whatever we output with this is aligned with that. It should feel AI. It should feel like it was created with humans in the loop. That is a faster, scalable approvals. We had to figure out an approval process, something that will really deliver on the promise of speed that this technology promised. And we were able to get that with this framework that I described. Like, Hey, produce something, do a legal check, make sure you’re promising not have any copyright in them. We will search those images and then send all of that context to your legal reviewer. So we would send out the output of the image, the actual prompt that we use. What do we get when we reverse search that image? And with all of those things in mind, is it good for clearance? 8 out of 10 times would be good, but those 2 out of 10 times, if it was flagged or rejected, guess what? We do not have to spend a week downloading a new image, really modifying the content and sending it out there. We can do it within seconds. So that was the first brush. Yes. With Jengai image integration, think about this as a crawl walk run. We were crawling with Jengai adoption and just what it can do. I was lucky to be in a place where we were all aligned across the organization and there was a need for this, a business need, a customer need and a tight enabling to make this happen. And that was the first brush. And with that, a new vision came to life. Now, remember how I mentioned that this kind of a presentation, we have work on personalization and really what is personalization? It is trying to show the right image, sorry, the right message to the right customer at the right time and see if you can move the needle. The right message in a digital environment often boils down to good things, your art and copy. And the example Jengai production that we just saw, that was largely speaking to the art. And if we can design one creator for millions, can we design for one server to millions and then flip it the other way around? Can we design for millions and then serve it to one user at the same time? We had just overcome production and budget limitations.

We could create visuals for, in theory, we could create visuals for every section of our audience. That is personalization at scale, the Holy grail all of us marketers are going to chase and decode.

And very quickly I realized, the same thing that we were able to do for creative production with technology. Now we need a data backbone to support that. I can’t have a solid thought of rules, but turns out I was unlocking personalization. Great job, Anish. But no, really the idea is if you can create those images with fast speed, do we have the right data infrastructure and the intelligence to then serve them with personalization at scale? So it just stopped there. Now that we could create those images, we now had a new use case and that’s really when Adobe AI Assisting seemed like that could help us solve for some of it. Now I am a simple-minded marketer. I am not a technology maker. And the CDP was extremely intimidating at first, because you’re talking about schemas, we’re talking about profiles, events, attributes, and it’s a lot of terms which, as someone who was just a copywriter, just making films, wasn’t really super comfortable with. So two years ago when I went to the Adobe Summit for the first time, they showcased what we know as the Genii check or Genii Assisting. And very quickly I realized, hey, this is something that can help democratize access to CDP. The way I was able to really get up to speed with CDP is we have a great team. We’ve got Kevin Mayfield on our marketing technology team, very knowledgeable guy, extremely open to share all the things that we know. And it was just me asking questions and spending time with him day in and day out. But there’s only so much time that both of us have. And if there was a way for everyone on our team or most people who want to understand this technology better to bypass that human interaction or just get enough of a common understanding before really making use of someone’s time, would that be possible? And that’s what Genii Assisting, the AEP platform, promised.

Some use cases that really come to mind are the knowledge navigator. Can we ask questions about the platform? Hey, what is a schema or what is an audience? How do you activate an audience in target? And also like a workflow accelerator, answering questions about business products. Like, hey, help me understand how many active audiences do I have? How many former profiles do I have in my CDP right now? These were usually questions that you have to think about, spend some time collecting answers, working with data warehouse or the CDP team to just understand. And these are simple questions. Can we bypass all of these fundamental questions which sometimes stop people from adopting the technology and get everyone up to speed? That was the vision that they promised. And really, what excites me about this is you can actually copy your data like team. I want to ask you hey, can you create an audience for me which tells me who will disconnect in the next four weeks? In theory, it should be possible, but I don’t think we are there yet right now. Or keep me honest, if you’re already using this, I would love to check with you how to do this better. But some of the ways we approach this right now is just very simple, formative questions which are at the start of any project discovery. Like, hey, what are lookalike audiences? Or how do I export gigasize? Very basic questions. Or how many active data flows are broken? Can we flag them early on and investigate a fix? So here are some of the prompts that if you guys have access to this tool should try out. Let us know how it goes. How many activator and streaming audiences exist? How many batch and activator audiences exist? How many duplicated audiences that are not activator exist? Really help you understand the health of your current platform and what can you do to help you understand the health of your platform. So personally, we started with video production or like we started with image production. Now as we move towards personalization, getting a good sense of understanding of your platform, what are audiences, what is possible with that. And when you join those dots, then you can work backwards and understand, hey, these are the five customer segments that we can explore. These seem extremely valuable. They have enough of an opportunity size. Now can we create imagery for them? And that’s the journey we are trying to decode right now. We have across the organization, across brand, across marketing, across e-commerce, across AEM. Do we have an understanding of how a CDP works? And once we do, what can we do with this great opportunity that lies ahead of us? So with that, we were able to get this cool feature in our work scheme and some inspiration to accelerate approvals for AI tools. Thankfully, the second time around, it was easier because now organizationally, we were more mature. We were able to tap into Adobe’s resources to understand and navigate some of the concerns. But really some inspirations for you guys to accelerate approvals will be anchor in business value and not the technology. Technology can very well run the risk of being a shiny object. But if you anchor the ask against business value, which is this much incremental revenue or this much incremental sales. So these are the number of hours we can save by incorporating this. You remember my first example? Genii can help you solve a lot of problems with speed, accuracy, and high quality and low price. Address risks. Come prepared with your data flow diagrams. Just really understand how this works. Do a lot of research. Speak with your legal team. When trying to adapt Genii technologies at scale, your legal team is going to be your new best friend. Trust me on this. Third is pilot first and scale later. Position the tool as a low risk pilot. Just try it out first. See if your hypothesis holds up. Another question that we are trying to solve for right now is, the CDP is extremely valuable. It can help us create lookalike audiences. It can also help us activate audiences directly to advertising platforms. How do we go about it? And it all comes down to, do we have enough business value? What kind of risks come with that? And then can we explore a pilot to solve for that? And recruit champions. The key thing for adoption is, for any technology, unique champions or early adopters to showcase the value of tech.

For Genii image production, I got a chance to be one of the early adopters and a champion of that technology. Genii Syskine, we’ve got a great marketing technology team to champion that need who helped the organization understand, hey, this is what is possible. Do we want to try and pilot this? Do we want to try and get this? And that’s where Adobe comes in clutch because Adobe also had beta program that we were able to be a part of so we could experience this first time. So getting those early champions, early evangelists in and giving them a sense of what is possible can help you accelerate that journey so much quickly. How do you navigate buy-ins and approvals for each stakeholders? Again, similar process. What does product care about? What does ID care about? What does opposite care about? What does marketing care about? Product cares about, hey, we need something that will scale with our data and enhance customer experiences. So we spoke about, hey, can we measure, can we have measurable time savings in building and launching personalized journeys? Can we go from looking at a week to answer some basic questions to having them on your fingertips? ID integration must be seamless with our current market. Agreed. We do not want to spend a lot of time integrating new technologies, which we don’t know if they will deliver the value that we have promised. Thankfully with Adobe, it was already integrated seamlessly. And when the team spoke to the Adobe business reps, it was easier to understand what it would really take to operationalize this. Operational securities, can we understand the security protocols? How is our data being managed? Again, helped by conversations with experts at Adobe, with our team understanding the technology better. So you really even see the parallel thing, how we approach the image integration thing first. Let’s get comfortable with the technology. Let’s understand what’s possible. Give you a first-hand experience of how this does and then see where it goes. And marketing, can it clearly drive a personalization efficiency and measurable brand impact? Can we do that? The approach is, hey, can we talk about some time savings that this can potentially unlock? And more importantly, especially for the journey I was asking, the CDP can run the CDP for being a niche technology, which very few people have an understanding of. But when you have an interactive interface, can you start talking with it to understand just what’s possible? Get a preliminary understanding. And if you guys have access to it, give it a try because it comes up with a lot of prompts that you can just keep clicking into and help you understand the platform better. Cool. Now you’re still probably thinking, where are the adoption tips? I’m going to summarize everything that I spoke about real quick before we pause for questions. The best practices for grabbing adoption of AI tools are, you have to keep practicing. You just have to keep using that technology. Keep pushing the envelopes to understand what’s possible today. You lean into your support groups. You’ve got the Adobe experience leagues. You’ve got a bunch of Adobe champions. You’ve got your customer success reps. Keep experimenting with those technologies and figure out what’s possible. Where are you hitting those roadblocks that stop you from adopting or even learning about this technology? Ask questions, keep playing with it. Join beta programs whenever possible to stay ahead. The Gen AI beta program was really cool and it was such a cool experience to be able to experience, to interact with a tool which wasn’t available for public release yet and to speak with the Adobe product teams that were making it happen. So we were able to give understand why is something not working or why is something working and really go about that.

Leverage AI to quickly grasp new concepts. That’s really what this whole Gen AI section was about. Understand with this technology, what are schemas, what are audiences, what is the general health of where the CVP is right now? And lastly, build communities of practice. Document and share prompt and best practices across your organization is another way. One other thing I really love about working at the place where we have a Gen AI practice group and communities where it’s just a large organization where people keep sharing some of the cool new AI tools they’ve discovered or trying to solve problems with the existing tools or just sharing knowledge and cool innovations that the whole team is doing. So it fosters the culture of curiosity and of adoption and really what do you need to grab adoption? You just need to keep using that technology. You need champions and spaces like these facilitate that. So in summary, share your guardrails first when dealing with any Gen AI technology. We just spoke about tool but the pace at which technology is moving, you got tools for any and every problem. Susan, you stole my answer when you said you want Gen AI to afford your loyalty. Mine too. I hate doing loyalty but we’re not talking about a guy who gives a case right now. Any problem that you’re looking to solve for, is it AI accelerated reporting, AI accelerated audience building, agentic AI? Share your guardrails first because the promise of technology doesn’t mean anything if it leaves your organization exposed to some of the shortcomings or lack of understanding around that. Showcase proof of business value, show clear time savings, accuracy gains and wider accessibility so every team on the right team can make the most of it. You have to build in scalability and security. Some teams or some products are extremely great with that. Even when we were exploring image production with Adobe, one of the biggest things was, hey this is legally not because the whole technology is trained on our own database and you have to account for those things when it comes to scalability, security, data security, compliance and you have to foster a culture of trust and speed. You have to move fast, you have to transfer, you have to verify.

When you are trying to do your pilot, early collaboration and repeated approval or approval workflows can let you innovate and do a lot more than we were able to do and we are trying to drive clear. So those are some final takeaways before I go into my what’s next in some of the things that I’m going to do and this is, I am a Genai evangelist and a really curious person and this is something that I’m trying to do on the side. I’m going to figure out if we can use Genai to create a personalized coloring book from idea from purchase to delivery in three weeks. What you see on your screen right now is a 100 step automation all mocore, agentic AI in nature and it’s something that I’m trying to do on the side and figure out, hey how far can this technology really go because if we are able to create tangible products by using soft data, by using digital data, that can unlock a lot of joy for people around the world. So the project is called Go Color Me. It’s an agentic AI exploration hobby which again was inspired at last year’s summit because I realized, hey agentic AI is so cool, what is agentic AI? A lot of times it is an LLM and with making the right API calls and automating that whole chain for us, that is cool. I want to try and learn that. So that’s one of the things that I’m exploring and I’m sure when you know if something comes off it, I’m going to be looking at the same challenges. What are the legal guardrails around this? How do I make sure the data is secure? And all the good stuff that I’ve been talking about. I’m also exploring a lot of cool things in the Genghis AI video stuff. Check out the resources section. It has the NFL spec ad that I just made. Something that looks like it could be half a million dollars in today’s production costs but a fraction of time and money to produce that. The technology is really accelerating at an unprecedented pace and it’s a great time to be working in this space. That’s all I have for us. Let’s get into the Q&A. Thank you.

Awesome. That was great Anish. Thank you so much for sharing. We do have a couple of questions here and a new poll. So please stay for the Q&A, answer our poll questions. The first one is from Becky. She’s asking for target testing, did you also measure a male basketball celebrity versus AI male? Female basketball celebrity versus AI female? We did a bunch of testing. This is again from a couple years ago. I do not remember if we did. But I’m sure we must have. We have a dedicated testing team which does an excellent job at really just understanding this whole problem holistically. Like I said in that section of the presentation, that was one of the tests that I could find at hand. But we found out that consistently these AI produced images with the right context can pull it up. And we are not always looking for conversion gains because think about all the things that I spoke about. We are able to save time. We are able to save money when we use these. Are they comparable in performance? And it’s got this guy help us offset some of that. Great. Pamela asked, and did you make the offers the same? The example used had different copy. Yes, we can use the same offers. The screenshots again, these are screenshots that hang. Great observation.

Awesome. And Heather asked, where do we find answers to the details that OpSec is likely to ask? Heather, this is a recorded session and we will be posting it on the community and we will send it out via email. Stay tuned for that and you will have access to the slides that Anish presented. Awesome. Keep the questions coming. There is a new one in here. In times when all enterprises are fighting for consumers attention and Gen AI is a common solution, what will happen when all adopt it? How would competition look like then? How do you anticipate in the next five years? That’s a great question. Something that I’ve been exploring a lot is what happens when the general benchmark of output is high and Gen AI is no longer a competitive advantage like it is right now. If you adopt a technology fast, you have a definite competitive advantage. But in a scenario like the person who asked the question mentions where everyone has adopted it at scale, what helps you break out and it all boils down to your taste, your timing and the tools. And these human elements and this is something I spoke about in one of our recent podcast episodes as well. Yaga farmers check it out. But when everyone is using technologies and efficiencies at scale, then it really boils down to your human elements which are, do you have the right taste? Do you have something which is of a specific, I call it taste but really can you make someone laugh in your unique way that only your brand can? Like look at Liquid Dick or look at Red Bull. Can you fascinate someone the way only you can? So that taste is a very human element. Timing. We recently had the first ever Gen AI created TV commercial go live on TV. A guy was this ad by a brand called Kalshi. It’s a betting app and they took a commercial on primetime TV, on ESPN for the NBA Finals. There was great timing. Timing again is a very human element. Can you leverage your tasting timings and the tools? Everyone will have access to these tools in the next few months or years as we keep doing more of these adoption conversations. As we learn more as practitioners, can you combine your taste which is a very human cultivated skill set? Can you combine that taste with timing and tools to create something which only your brand or only your business can? So the humans in the loop for a big competitive edge. Long-winded way of saying hey, people are awesome. Awesome. Thank you Anish. We do have a couple more questions. Derek just added, what do you see the long game begging with AI? Long game. It’s accelerating or allowing us to bypass a lot of challenges that we face today. Let’s talk about one example. It’s a very broad strokes question but because I have a creative background, I’ll speak to the creative piece of it. A few years ago, who would control the creative? The clients or the companies with the right budgets who control the creative. But if you can now create a high quality video, high personalize to each customer segment, suddenly budget is no longer a barrier. Long term, I see AI just accelerating access and democratizing what’s possible for everyone. Look at the coloring book example that I shared. I’m not a coder. I am a filmmaker or I’m a personalization practitioner now. But with the help of these mocore tools, I could create a SaaS product which wouldn’t have been possible before the age of AI. It’s really opening access up to people who have ideas and now taking those barriers to execute those ideas away. Awesome. Thanks Anish. We did have another question from the chat. What advice would you give a marketer just getting started with AI tools? Keep using them. Just be curious. There’s so many tools coming up every day. Majority of them are going to be noise. Majority of them are going to be chat, GPD or LLM wrappers. Things you can do natively in an LLM like Gemini or Perplex City or chat, GPD. So really understand what does technology do? What is it that you need to do? Start with a use case. Start with a problem statement. Susan, I love that problem statement that you started with. Can AI afford my laundry? But we are not there yet. But the use case that I started with was, we need high quality images. We need them fast. What’s the best way to get them? Start with a very real problem. There may be an AI solution that exists that can help you get closer to solving that problem and keep exploring. Stay curious. Awesome. Christina, you’re asking, what is the legal terms that protect the companies generating AI on Adobe on any copyright issues? I did link out to a security fact sheet in the resources box. I would highly recommend reviewing that. I think that will help answer some questions. Awesome. And then keep the questions coming. There’s one more in here. Is there a requirement to label AI generated images in social media creatives? Depends on the use case, right? Like if you are using like a lot of it also drives from what are your ethical and companies legal stance. Ethically, you should not label anything that misrepresents a human in a, like when you’re giving a product testimonial, right? When you’re giving something like a product testimonial, you do not want to misrepresent the fact that this is a real person. So that’s a very clear memo. One of the most worrying trends that I see personally in the Gen AI video production space is user-generated content, which was generated artificially. And that looks really, really good and can be so misleading.

So lean into your ethical and moral responsibilities and label them as needed. Yeah. Awesome. Well, I think that’s a wrap for our Q&A. Thank you all so much for joining. We, like I mentioned, we’ll be sharing the recording and the slides after the webinar. And if you don’t mind giving us a quick poll check over here, we’d love to hear what AI topics you would like to see us cover in the future. And if you’d be interested in sharing your story, so please let us know and keep them coming. Thank you all so much.

Thank you.

Session details:

  • Creative-first AI thinking: How framing AI as a partner shifted internal perception and builds trust with legal, security, and business stakeholders.
  • Frameworks over features: How focusing on decision principles not tools enabled safer AI adoption.
  • Learnings: Real-world insights and reflections to help your teams avoid common pitfalls and accelerate enterprise AI adoption.
recommendation-more-help
76e7e5df-e11d-4879-957b-210b1656662e