044. The Impact of Artificial Intelligence on Digital Advertising

Foreign this is Emilio Garcia from Mountify. Welcome to another episode of Demand Gen Studio. With me, as always, is Gabby, Gabby Garcia. How are you, Gabby?

Hi Emilio. Feel fine. How are you?

I'm doing great, I'm doing great. I'm hoping that everyone has having a great day for us. Today is Friday in the morning, so really excited to end the week and what better way to do so than talking about this for sure hot topic which is AI and the impact on digital advertising. I we have been, you know, Gavi and I talking about this for a little while and we, we thought appropriate to, to have a conversation about it. For those of you that you know, might be listening to these down the road, just follow along on the topics themselves. But today we're going to experiment a little bit with the presentation and so we have the, we have like a little presentation here to guide us about it. And so those of you watching in watching might find it useful. So what we want to cover today obviously is, you know, I find hard that anyone listening here or watching will not know what AI is but still cover a little bit on that and some of the general benefits of using it for digital marketing. Of course, AI will have many applications for a lot of industries and it has been having that impact for years already. To be honest, it's not new. We, at least in the marketing space have seen the impact of AI on advertising for years already and big platforms like Google, Facebook, obviously Microsoft have been using it to some extent already for quite a while. However, given the visibility of ChatGPT, it is increasingly evident that this is taking a next step and more people is becoming aware of it and is getting definitely traction. So we want to explore some of the immediate impacts and maybe explore as well some of the things that will evolve around it. So that will be the kind of the agenda for today. Obviously, very generally brief description of AI is a branch of computer science that deal with the creation of intelligent agents that to some extent can reason, learn and act autonomously. Right. And this is something that has been growing quite fast. Obviously I'm not an expert on this, but the general understanding that I have of it is that similar to a neural network, like the way that our brains work, you have small pieces, in this case in the form of software that can be rewarded for a specific outcome. And instead of describing a specific set of steps, you supply data and develop a mechanism to reward those small pieces of the whole model. And then through connection, similar to what the neurons work, some positive feedback is strengthened and things are not in the right direction, diminishing value and a network is created and very similar to the way that a child learns. Right. So the greatest takeaway from this is that while the goal of the trainers and the data set is known, the properties of the agent emerge from the training. And that's for me the biggest, you know, like or the, the biggest impact of how AI is different from regular, you know, computing activities like programming where the, the programmer already knows the outcome and how it works because they have detail step by step how the output is going to be created or calculated. It's different in AI because just like a, just like with, with a child, we teach them stuff and, and they learn it but we don't have clear understanding of how those connections, how that learning process went by. We can only see from the exteriors, exterior and we, we don't have a clear understanding of how those connections are built. And that's important because every time that a new AI is trained, new behaviors will emerge or surface and the only way to discover them is similar to what we, how we interact with intelligent agents, humans through questions to back and forth through or through communication. And I think that's very exciting. That will obviously open up a lot of, you know, interesting venues for what is now called prompt engineers where depending what they're asked, they will discover the capabilities of these agents. And I think just obviously it's going to get more and more powerful over time. Now however, obviously in our space right now for demand generation professionals and for digital marketing in particular, we want to know well, what are the specific impacts that I can expect from AI for digital marketing. And we have, you know, obviously there's a more extensive list but at least we see that the best areas or the highest areas for impact will be personalization. Right. We can use AI to cater more specifically to our buyers and to the type of companies that want to reach with individual message that might be composed in real time. So that's, that's important because right now requires a ton of work from us to make it happen and sometimes it's not feasible economically. The next one obviously is content creation. And again here probably is more for the most grueling part of the work which is the research and getting the references. There is probably in content creation the biggest area for doing mediocre work because people might try just to compile what is already out there from the data that is feed but obviously the proper use has greater implications because will allow people for example to transform into alright content contact context, sorry, content created from e.g. thought leaders that is hard to transform into text by themselves or it takes time from the creative team. So I think that area is going to see a lot of, you know, progression over the next few months and years. Obviously the whole point, not the whole point, but another big area is optimizing performance and especially campaign optimization, which this is the area for example that I can tell you for years has been already working as, you know, all those bidding strategies that rely on data like maximizing conversions or maximizing return ad spend. They use machine learning to feed the algorithm the reward. And after exploring for keywords, ad combinations, demographics, locations, that is transformed into a signal that, that allows campaigns to optimize and that would just continue. We continue to see how it has impacting, you know, testing multiple headlines, testing multiple descriptions. I am sure that variations on images will be added to top of that. And obviously the rewards themselves will evolve. Not just a specific event like signaling a conversion through a thank you page, but exploring the whole website or exploring all the signals that might come from the CRM. Eventually I'm sure are going to be feedback into campaigns to optimize in ways that probably we humans will not understand completely, but the AI will understand pretty well. And finally a little bit of fraud detection, which is it over itself, right? I think it's going to be an ARM raise. We have used some tools already for fraud detection that pretty much understand, right. If the agents clicking on those ads or engaging them are real humans or our other agents or AI. So it's going to be fun to watch that, right? Using AI to detect AI. So we prevent us from ourselves from AI. So it's kind of like a circular reference there, but it's happening already, right? We have used some tools that do that and I think are going to get more powerful because they will have to get more powerful because there are more people that is trying to fraud systems to take advantage and you know, divert investment from advertising. So those are some of the areas and obviously the benefits will be related to if we can create content that caters to a more specific demographic, then they will find that content more interesting, they will engage more with it. And to some extent, while it is impossible to measure everything, we are measuring more and more. I think that's the balance that we want to strike there, right. So obviously there's privacy and we want to make sure that we are not identify a single individual in our measurement. But I'm sure that tools will be developed to have the best of the two worlds as much as possible, right? Being able to protect privacy, but at the same time being able to identify behaviors, patterns, engagement as much as possible to allow people to optimize toward their goals. So, so those are some of the areas where they can be used. Obviously specific areas where the AI is being used. For example, in the case of social media ads, specifically on social media ads, I think the biggest ones will be around engagements and behaviors. I think so far, for example Facebook or LinkedIn has been building these static definitions of behaviors. There are classifications and I think more and more those definitions will come from maybe phrases or keywords and that will compile a dynamic audience of sense. Right. So for example, some ABM tools use, you know, have organized historically in 10 topics. As that topics classifications made by either computers or humans are restricted. But more and more we are seeing that now those intent topics are becoming pretty much keyword topics and you just come up with the keywords that you feel identified so you have way more flexibility and probably those language models will match those keywords to whatever signals match to that idea. So I think that's where there's going to be a lot of progress. Again obviously social media being so heavy on image and on video. The two components that will add a lot of effort here or reduce a lot of time worth are the creation of the ads themselves. And then once you have them running, the other big one will be the variation of them. Right. If I want to make a slight changes on the fly, I think that's where AI can support. Not because we want to find a better version overall, but, but because we want to find the version that caters to specific person the best. Right. So more and more I believe that there's not necessarily or in the future there's not going to be something like the best ad or the best headline overall, even when you. We can cover a big bunch of people over idea. But I think more and more we will be finding what, what interests a particular person or a smaller group of people and that's what we will deliver. And the only way to scale that will be probably through AI to be able to do it, you know, in a very granular way, but especially in real time, which is the part that is the hardest. So those are the, the, the ones that I feel are going to impact there now in search engine obviously again here the whole idea of indexing is something that came with a lot of rules and again this is an area that most people are not aware. But Google has been using AI to organize content and discover behavior for years already. Right. More and more of the results that we see on the web or for ourselves are just for us. So you are not optimizing anymore. Very similar to what I was saying about social ads. My first result will be different than yours because the system is learning your patterns, your behaviors, it has your history and then is changing the way presents information to cater to what you are interested on, which is good and bad. Right. Obviously that one of the of the problems of that be of that idea is that then we build up more or strand our, you know, make our biases stronger, which is not good. So there's obviously some work to do there to make sure that we have a more balanced view of the world through search. But that's the area where most of the implementation is going on. I have seen that for example, more and more the engines do not use specific areas of the page as the result, right? Like the headline and the description to put what you're going to put on the snippet. More and more the the machine is parsing the content itself and coming up with summary of source, if you want to call it that way that better inform people what is the content or the page about. And it's going to continue to improve upon parsing audio, parsing video. We have seen that now, you know, in the past YouTube videos used to require or still do and require humans to define chapters and sections of interest to break it down. And more and more we are seeing that some p. Some videos are being break down by AI to discover the sections that matters. And the same, I'm sure is going to happen to audio in similar situations. And finally, specifically on advertising we see that more and more we are going away from, you know, specific keywords having, you know, exact match and frank phrase match. Those are less and less as restrictive and broad. Match definitely goes into many, many areas. There's still opportunity there in the way that search engines understand, you know, potential related results that are the result of, you know, not being in the real world, I guess. But I am sure that it's just a matter of time where we probably will, you know, get away from very specific keywords and it will be more about. This is the offer that I have in the form of a landing page and these are my main, you know, topics and Google will find the right searchers for you as long as you have a reward for it, which will be the conversion. So more and more moving into dynamic type of campaigns for us, for those of us that have been working on this space for many years, we see how we have less and less control over the targeting. And so I think that will continue. There's going to be still obviously opportunities for those that understand how that targeting is set up to try to improve and take advantage of the things that I cannot do very well. But obviously I think it's just a matter of time before we should focus more on the, you know, the basics of marketing, which are who is the audience that we want to target and what is, what is that we are and want to say and how we are different. Those kind of, you know, more strategic decisions. So that's there on the display side, very similar to social media. I feel here that the biggest opportunity is how to determine where do you show your ads based on the content of the website or the user behavior. Right. So that's the biggest opportunity. And here obviously the big. Another big problem for display in general is fraud. Right. So I think AI here will have the computer power to discover more frequently and faster websites that might not be the right choice for the ads that we have in mind. Testing variations because usually display have so many sizes and formats and so if I can have one version of the ad and in real time being, you know, create variations on it, that will be fantastic. So being able to select, you know, between text and image and audio, the right format, I think those are going to be interesting spaces where display should improve. Especially display on mobile where most of the, you know, if not necessarily fraud, at least wasteful impressions and wasteful clicks through that coming from gaming should improve. So I'm excited about that area in particular. Obviously landing pages, right. The offers themselves. This has been evolving in many, many ways. Starting from more and more being able to create layouts that are effective quickly without the need of, you know, the involvement, heavily involvement of designer. That's an area that I think is going to have great impact. Obviously the ability to come up with, with copy and you know, interesting headlines and descriptions. We were seeing that HubSpot is launching content assistant that will allow people to get ideas for the headlines and descriptions they want. Obviously will be fantastic if for example, the tool will parse your website and when you're building that landing page based on the content that you already have, will make suggestions about what you can do different and, and that is still aligned all the colors, the typography that can be sourced from the website that you already have. I think those areas are going to save us time tremendously. Right. And finally once you launch them, the ability to make variations in real time, test ideas on their own and find winners. I think that's also was going to be helpful. A B Testing is an activity that consumes a lot of time, requires a lot of understanding of technicals and tools that are able to create meaningful tests. Right. Understand where it makes sense to do some will be, you know, a great add to marketers in general for sure. Another big area, I think that's the last one will be how we will use AI to find, you know, the right keywords or demographics on search campaigns and social media ads. Again, going more and more toward the intent and understanding the buyer and the ICP rather than just the specific keyword. I think the understanding meaning is one of the biggest challenges that marketers have when they are trying to come up with a list of ideas for keywords or understanding, you know, the motivations and behaviors of buyers when they're trying to come up with a list of demographics and firmographics for their campaigns. So this is an area where definitely, you know, AI will contribute a lot. So in general. Right. Those are the areas that we see are going to be impacted by AI. Finally, probably just some of the trends that we see in the future that we have covered already. Right. More personalized marketing I think will continue toward that, you know, like a topic future where an ad is built just for you, just for your eyes. I think that's definitely something that's going to happen. Intelligent content creation. Right. We need to move from reusing what is already there to coming up with some things truly new. I think that will be the barrier to, to move toward. I don't know when it's going to be, you know, accomplished, but I think that will be the next stage obviously on the marketing side and the automation. Building automations is something that is very time consuming as well. So if AI can start coming up with best practices on automation path, discovering where automation is required, which is the other big problem. Right. I think that will be fantastic. We have talked about real time marketing, right. Changing what we put on the ads as users behave based on their engagement. And in general, marketing platforms getting more and more help from AI to deliver better results. We already seen it on most ad platforms on CRMs and marketing automation tools. So I think we'll just continue and evolve over time for sure. So that's what we have for today. I hope you have found value in this. I don't know if. Gabby, you want to add something?

Sure, yeah. This is a like a note to the, to the main part of this conversation which was that AI is growing like really fast in the, in past years has Been like a, like a trending topic, but now it's just like it's going to be overwhelming like the amount of people that are interested in AI and the amount of money that has been invested in, in it too. I remember telling you that a couple of years ago, like in 2021, the global spending on AI was about 50 million and it was expected to be at least a hundred million by 2025. So it's going to be like really like, like a boom, right? It's something that everyone's interested right now. There's a couple of professions that are racing because of this. Actually part of my job on Bondify, it's about using AI in my daily basis. And actually I wanted to show you like a, like a side of it, like what can you do on, on your daily basis as a marketer? Someone that creates content. That, which is the part that I use for it and, and let me share it if I, if I go.

Go for it.

Yeah. This is like just a site, right? About creating a post that you can, you can use on, on LinkedIn, you can use on Facebook and you have to use prompts. The way that AI works is that they, they learn how to communicate what you're trying to reach, right? So you create a prompt and it, it starts analyzing it and wherever you are being more specific about it, it starts to giving different answers and it's going to, it's going to be try, it's going to try to be more precise in, in the way that it's using that information that you're giving to the AI. So that's like a side, like a, just a pinch of what you can do on AI and, and the content you can create. If you can see on my, on my screen there was a lot of things that I used for, because you create chats and you create projects through AI just for any single thing that you, that you want to accomplish as a content marketer or as a content creator. And that's just a pinto you can do. Right. If I knew how to show you all the things that you can do even on Google Ads and the way that the AI works, like the algorithm works and how it performs every single time, like, like better. There's another. Right? So there's a lot of things that AI can help you to accomplish your daily basis objectives as a marketer. So yeah, I wanted to share that.

Yeah. A prom engineer in the making. So yeah. And I think we will be sharing more and more content about this. More practical examples. We have launched A series of webinars that are just gonna, they don't have, they are not going to have the frequency that we have for the Manchin studio that will be, you know, know every month. But I think we will have opportunity and Gavi, we will take advantage of your, you know, increasing knowledge on AI, how to do things. Obviously, how do you use Chat GPT, but it's not the only AI tool in town, right? We have order tools. Like, I've been dabbling a little bit with copy AI and similar tools. So I think we will share more and more how you can do that on your own content. For sure, sure.

There's bar than copier. That's another one that you can use. Actually. I told you that Bart is. It's. It's the Google one, the ChatGPT for Google. And, and it's really precise, like the information that it gives you. I, I will say that because of, it's the main search, you know, like the, the main place that you go for a search. It has all the information that you want. And, and the only thing about Bard is that you cannot save the content that you actually are trying to figure it out. Like, you cannot save it to projects or to chats through the conversation with the AI. So that's the only thing that, I mean, there's an improvement, like a, like something they can do, they can do better on that topic. But yes, and it's really, really important to learn now as a marketer how to use AI, because like we have been saying, like I joke, you are not going to be replaced by an AI, but you're going to be replaced for someone that knows how to use AI. So that's something to keep in mind.

I agree. That's exactly what's going to happen. It's just like the joke about the bear, right? It's like, oh, you don't run faster than a bear. Yeah, but I run faster than you. That's all that needs to happen. So. All right. Well, Gabby, thank you as always for joining us. I hope that you have found value. We will continue the conversation about this and until then, have a great day.

Have a great day, everyone. See you next time.

Bye. Bye. Bye.