The Business of Digital: Episode 37 - AI Transformation in Media

Speaker A: Foreign.

Speaker B: Computer.

Speaker A: This is a multimedia system.

Speaker C: Welcome to another episode of the Business of Digital podcast. Today we're talking about AI transformation and not the theory, but what it looks like when organizations actually start putting AI into action. I, I'm joined by three leaders who are right in the middle of that change. We have Noel Geer, Founder and CEO of Connected Interactive. Sahil Razdan, Senior Legal Counsel, Samsung Electronics Canada and Samsung Ads Canada, and Philip Kleim, Vice President, National Revenue and Digital Advertising Strategy at Chorus Entertainment. Thank you for coming over. So before we begin, uh, I'd like you guys to uh, introduce yourselves.

Speaker D: I'm Philippe Kleim, Uh, I work for Chorus. It's one of the largest media company in the Canadian market. Um, you've probably seen some of our content. We have brands like Global tv, Global News, um, Flavor, Home Slice, Adult Swim and many, many more. Uh, what I do at Chorus is I always see a sales team, um, that is responsible to sell our linear TV and our digital assets to agencies and clients direct.

Speaker B: Noel Geer, uh, CEO and co founder of Connected Interactive. Uh, a little bit about me. I've been in the advertising industry for 25 years now. Um, I started my company in 2012, focused uh, originally on really the mobile space and understanding, you know, how the app market plays with online. Um, moving into programmatic and have been really in the, in the, in the space of programmatic advertising technology and data now for the last 10 years building audiences Connecteds. What Kynected is best known for is our ability to bring in offline purchase data into advertising capabilities and bonding offline behaviors with online behaviors so that you can target users based on actual purchases, um, that are happening by both in store and online. Um, because today about 84% of retail, uh, sales are still happening in store. So it's a really big portion of what's happening. And uh, we've partnered up with the likes of Baneras, MasterCard and so forth. So in Canada specifically we see about half of transactions in the country, uh, that we could map to audiences and so forth.

Speaker A: Okay. Sahil Razdan, senior legal counsel at Samsung Electronics Canada, um, do a lot of their data privacy work, data protection, AI type work. Um, biggest internal client is Samsung Ads Ads Canada, where we do anything from out of home, um, ads to website ads and any other, um, campaigns that we might have managed from the legal side there.

Speaker C: The three of you are also chairs of our respective committees. Like Sahil, you're chair of the AI committee. Noel you're with the mobile committee and also Philippe, you're with the CTV committee. Uh, thank you again for being here, uh, for this podcast. Now, let's start with you, Noel. Um, you've been leading what sounds like a full scale AI transformation, like retraining your teams, rebuilding your tech stack, and embedding AI across your operations in your tech. Uh, can you tell us what that journey has been like?

Speaker B: Absolutely. Uh, we knew this was coming a long time ago and, um, know there's, there's, there's no question, uh, right now is the opportunity to move on this, um, because it's not going to be too long from now where it'll be, you know, too late and you'll just be kind of moving the way everybody else is. So we saw that, We've, we've obviously seen that coming for a while. But the biggest thing when you're in a company that you know is busy on an ongoing basis is where do you start? And so for us, starting with the fundamentals, uh, setting up education, um, spending the money to get the resources in place to start building on the fundamentals internally, so investing into the education piece so that everybody in the company actually understands what this means, um, and starting there so that as we sort of start to build out, you know, workflows and thinking about, you know, agents, synergy, AI and all that wonderful stuff, the fundamentals are there. So everybody understands the why, um, versus just trying to shoehorn, um, AI into the business. And so that's really where we've started. Um, and now that transformation, it includes building. So we've been building technology now since 2012. I mean, we built our first mobile SDK in 2014. We were building dashboards, platforms, all kinds of stuff, um, that proprietary technology. Our position now is really, uh, we're moving away from the proprietary tech. We've got amazing data. That's what we're really diving into. And we're going to the cloud, we're going and building the tech stack so that the tech stack rises as the technology rises and we can take advantage of all the coming changes that are coming in the market and stitch together in ways that are proprietary to us. So, so thinking about how AI, because everything's going to change so quickly, we want to be in a position where our data is the unique standpoint, how that gets integrated into these systems. Right? So that when all these things change, we're just changing with it. When new AI systems come into place that we can take advantage of, we're there and we're rising with the tide. Um, but we're finding unique ways to use it. And that's really been the key for us. So, so anything from internal workflows, uh, to even you know, getting into systems that will automate for us. And we're talking from you know, running campaigns to optimizing campaigns, to optimizing automation within operations, uh, to even building proposals, doing research and even little you know, once the data is flowing, using things like precision AI where you can build scoreboards on your business and understand really finite unique things on a daily basis that are happening. And that's really where we're going.

Speaker C: Philippe, how does that compare to what, what is happening in a uh, media organization like yours?

Speaker D: For us at cars, there's, because we such a large company, there is many, many different use cases where I can be leveraged. Uh, so I think for us um, the way we thinking about it, and I'm going to separate the two pieces. One, um, more from a corporate standpoint and one more from a view of what I do on a day to day basis which is advertising.

Speaker C: Right.

Speaker D: Uh, from a corporate standpoint we have an AI committee. Um, it's a group of people that come from different parts of the organization. Um, sales, legal, content, um, pretty much everywhere. Um, and each and every one of those teams, they come up and they submit use cases for AI. Uh, and it could be uh, I want to automate the way we look at content. Uh, I want to look at different ways of reviewing legal paperwork. Uh, I want to look at data assets and how we treat that. And then the AI company uh, looks at it and basically decides what is required, um, uh to get it done, uh, and if the use case is even viable, uh, on the advertising side I think what we really have been thinking about, I would separate it into two buckets. One that is more around what we want to build internally uh, using AI. And that's a lot, um, very similar to what you said Noel, around how do we optimize workflows, how do we optimize processes, how can we do things quicker? And in that area I would say the stuff that we likely going to build and the stuff that we're going to rely on external vendors to help us with.

Speaker C: Right.

Speaker D: And then there's a whole other facet uh, of how in my opinion AI is really going to be leveraged is everything around the data capabilities. Right, right. How is data being utilized? Is there an easier way, uh, with artificial intelligence to better look um, at this data and uh, activate on that data. And then the last thing I want to say is A lot of what's happening from a sales perspective is also working very closely in collaboration, uh, with the agencies and the clients to kind of understand what that AI connectivity looks like between all the different partners. It's an interesting chapter for us. I think being a big media company, there's a, there's lots of things that we have to be careful with, uh, especially when it comes down to privacy, when it comes down to the trust that you put into AI agents, etc. Uh, I would say we are very, we are the beginning of that whole journey. Um, and, um, to, to your point, I think our AI, our AI approach is going to evolve as the market evolves, like, and the pace that it's going to evolve at.

Speaker C: It's quite interesting because AI, it's like a buzzword for a couple of years now, but seeing the actual practical use being done within your organizations makes it so that, uh, it's coming into reality. Uh, what about you, Sehel? What's your experience?

Speaker A: Yeah, so, I mean, uh, everything that you guys said is exactly on point and from a little bit of a larger scale. Once you start implementing it the correct way and when things get integrated together, you see things like, um, the way we can integrate, integrate your, your watch to your phone, to your tv, to your dishwasher, to your laundry machine and things like that. And, you know, the AI capability allows you to do all of that on such a large scale and create this environment within your home that you can, you can leverage and you can make your day so much easier and your life more, more efficient. I mean, there's, you know, a random example is, um, you know, if you're, if you're doing laundry, you know, a lot of the time you do your laundry and, you know, you don't want it to sit there once it's done because then it starts getting that, that damp and that stale smell. Um, but, you know, you can do something like throw your laundry in the, in the laundry, um, leave your house when you're 20 minutes, 30 minutes away from home, go on your watch or on your phone, turn your laundry on, so by the time you get home, it's ready to go. And then, you know, like, if you're, if you're sitting in a different room, for example, you're sitting around watching tv, your laundry's in a basement or at a different level of your home. When the laundry is done, you'll get a little notification on your TV or on your phone or on your watch. You can set it up that way. So, um, to create just efficiencies within your life. Implementing AI can help that in numerous ways for sure.

Speaker C: Your work really lives within the intersection of mobile, uh, CTV and AI. How does that ecosystem operate on the ground?

Speaker A: Yeah, so I mean I gave a few examples of how that works just from a kind of granular level, but the way that you can incorporate that into a day to day lifestyle and the way that that kind of incorporates into, into ads and marketing and things like that from uh, from a consumer level is um, you know, you can, you can understand with AI, uh, you, you end up understanding behaviors and you understand work, work, uh, work life balance and work styles and behaviors and things like that. Um, so with the, with the different pieces of technology that you have, for example, let's say you have a watch and you work out all the time, you use the, the, the workout apps and things like that. M. You can set it as precisely or as non precisely as you want to eventually get ads that are targeted to your lifestyle of working out. Um, now if you transition from you know, being um, a single person to a married person then to eventually having kids, um, the searches that you have on your phone or on, on your, on your laptop where you know, you're searching up how to be a good dad or you know, what's, what to, what to do, like what to expect when you're expecting all those things kind of filter into the uh, the ecosystem and into the diagnostics of everything and how you interact with your, with your devices. And of course you know, you can, you can not choose to do that if you want to. And if you want to, if you want to, if you don't want to get these targeted ads and these specialized ads and marketing tool from your uh, to, to your devices. But if you choose to incorporate those things into your life, there's a huge efficiency that, that comes along with that if you use it the right way. And um, you know, it grows with you as you said, you know, you create a tech stack and you know that continues to grow as things evolve. And it's the same way with AI and especially if you're integrating all of your, your lifestyle, um, Technology so your TVs, your phones, your watch, it grows with you. So when you transition through these different phases of life and you search different, you know, keywords and criterias, or you interact with different ads, or you purchase different things around, you know, within, within different, um, different phases of your life, the AI grows with you. So it'll also incorporate giving you potential examples or suggestions as to what to buy next or what to look at next or a good book to read or what's good for food for a baby or you know, what's good for a healthy protein intake and things like that. So incorporating all of that together um, is definitely again brings so much efficiency to your life and grows with you as you continue to grow.

Speaker C: I guess what I want to follow up with that is that you know, Philippe, from your end, what does that look like when you're a traditional broadcaster entering the space? A very similar fashion.

Speaker D: So I think uh, you talked about it uh, quite the right way. Um, and I'm going to add. So you, your previous question was about AI is a big topic right now and everybody's talking about it. So we've had quite a lot of conversation and I had many discussions with many partners and, and clients and even internally with people that I work with on a day to day basis. Um, I think for us I think there's really two things that AI can do really well. Like one is to ah, your point, the connectivity between all the different data points that it's becoming harder and harder to manage especially when you are really large business. Um, and um, it's not just managing the data, it's also getting outcomes and information from those data points that you can actually leverage uh, in some, in some shape or form. So that's, that's the first thing. The second thing you can do everything faster with AI. And I had a conversation with somebody recently and he was saying um, what we were doing in five or six or eight months, we now do it in three weeks. What we were doing in, in, in, in a month, we now do it in a couple of days.

Speaker C: Right.

Speaker D: And it's if you structure yourself properly. I see. I think this is, this is what we are working on really as, as being a, a uh, large media company is, is really structuring yourself to achieve those two goals is make your um, your information cleaner and then make it so that you can go faster in making decisions, making faster in terms of like dashboarding and reporting and, and all the different things that uh, require human uh, interaction. But if you really want to do everything manually, it takes you a lot of time and AI can really simplify a lot of that.

Speaker C: Right.

Speaker D: Um, and I think there's many use cases for this. Like one of the use cases that we've been working on for example is uh, quite often we are asked about contextual targeting. Right. And contextual targeting. Just imagine when you think about connected TV or even linear TV all the shows that we run on a daily basis, like every single moment in that video content, how do you actually structure it and segment and properly look at it? So this is a use case, for example, where AI could help us look at all the different moments and kind of figure out what those moments are so that we can better help our advertisers and agencies figure out where they want to be placed.

Speaker C: And I'm also assuming that this is happening within still with respect to user

Speaker D: privacy, being on the legal side is how do you stay privacy friendly and how do you make sure, um, that whatever we do, um, there is uh, a, uh, clear way of keeping our data protected, our users protected. Uh, and I think a lot of the time also when people talk about AI, it's that trust factor. Right. Like how do you make sure that whatever you do within that AI ecosystem helps you, helps the consumer, helps your clients, but in a way that it's completely safe and trusted? I think that's, that's really the piece that is important.

Speaker C: Right. Anything to add to that?

Speaker B: No, so much. Yeah, uh, there's a couple things. I mean, just to kind of touch on the speed factor. You know, we always look at everything from the viewpoint of value.

Speaker A: Right.

Speaker B: And value, value is really, it's what are the outcomes you're trying to achieve, um, over top of the time that it takes to get there and the pain that you have to go through to get there. Right. So I'm trying to, you know, goals and outcomes over top of time and pain. And what's interesting is so often when people talk about value, they always talk about the two top parts.

Speaker C: Mhm.

Speaker B: But the actual unlimited value is speed and pain. So. So when you can get their fat, people pay more for speed. Ozempic is so popular because people want to get rid of fat quickly and they're willing to pay for it versus go to the gym for eight months. The unlimited opportunity with AI is that speed component and that creates such massive value. The other part is the pain that it takes to learn to get to a certain point.

Speaker A: Right.

Speaker B: And so when you can take speed and multiply that, um, exponentially. And also the opportunity with AI to use things like synthetic data to test and learn before you even go to market and actually avoid some of those pain points really is an exponential growth on value. So the ability to get somewhere quickly, test and learn in an environment that actually doesn't cause so much pain to get there is massive in terms of value. Now on the privacy side, like that's all I Mean, we are, we are as far as data is concerned. You know, we're holding a database of 45,000 different audience behaviors and we have to protect that. And it's really important. Privacy first always is the motto. Because if we don't have that, then, you know, we can't do what we do. Uh, and I think that that's also the part that we have to really, really hone in on. You know, so when we're talking about it, trust is the big component. And that's where I think the technology still needs some work. So the web itself is built on trust. There's the flip side to AI where that's actually in danger long term because can I trust the information I'm seeing? Do I know that that was actually a real person?

Speaker C: Right.

Speaker B: The flip side of that is that is that trust component. That's where I think the opportunity for things like blockchain and web 3 come into play. And when we think about how do we maintain that trust, we think about decentralized solutions like the blockchain, like what IB Tech Labs is doing right now to give publishers the ability to, to look at this stuff, have tokens, verify and create a verified system.

Speaker C: Right.

Speaker B: That's the part I'm really interested to see where that goes because I really feel in the next two to three years there needs to be some, you know, uh, stakeholders involved pushing this as sort of the next stage of where we're going. Because that, that trust is so important.

Speaker C: Yeah, absolutely. Now I want to go back because you did mention a lot about what you do within Connected Interactive to implement AI, uh, across the organization isn' Those are great examples. But I want to get into the reality of adoption. Uh, I think in one of our conversations you did mention that a huge number of AI, uh, implementations would fail at around a rate of 85%. You know, you can check me on the data. Um, but why is that happening?

Speaker B: So there's a stat that, you know, is out there that just in general, on average, um, there is, you know, implementation happening. And it more often than not, it's not giving people the outcomes that they were expecting. That that is because of the shoehorning of using.

Speaker C: They're forcing AI into the process.

Speaker B: Yeah, that's not so, that's not to say that there's not a huge lift. On the flip side, there's a 70, 75% lift on, on efficiencies that people are seeing.

Speaker C: Right.

Speaker B: But again, it's the speed and pain. So testing it and learning it over and over again to get to that point. So it's the number of implementations to get to the point where it works. That's where the fundamentals come in. And the fundamentals are such things like prompt engineering.

Speaker C: Right.

Speaker B: That is one of the biggest differences that I've seen in terms of making it give you the output that you're looking for if you don't understand the fundamentals of how to build a proper prompt. And I got. Sometimes it takes forever to build some of these prompts.

Speaker D: Yeah.

Speaker B: Well, you want to, you want to supply examples, you want to supply specifics, you want to. And the craziest thing that I've seen is the fact that if you tell the AI how important this is to you, it will actually improve the outcome.

Speaker C: Right.

Speaker B: I mean that is pretty interesting. I had a, you know, I built a prompt to help my mom search for something and I told the AI this is for my 85 year old mother and she really needs this. And they're seeing evidence now that these types of like notions, emotional prompts within the prompt.

Speaker C: Right.

Speaker B: Are actually increasing outcomes by 15 to 20%. There's a platform, um, called PromptMetheus.

Speaker C: Yeah.

Speaker B: Where so there's different sections of the prompt. Right. You got your role, uh, you got your task, your specifics, you got your examples, you got your notes. And these are, you know, use hashtags to kind of create what they call markdowns so the AI can understand what each section. So there's a platform called Prometheus. You can put it into each section and actually test the prompt and test different LLMs to tell you how expensive will be to get the outcome from the prompt. But it's like a, it's like a prompt optimizer. It's unbelievable.

Speaker D: Yeah.

Speaker B: Uh, but things like these and implementing this, but in order to even be able to think along those lines, the fundamentals is, is extremely important.

Speaker C: And I think aside from the fundamentals, it's also governance, um, within the AI, transformation projects within each and every organization. Right.

Speaker B: Um, yeah. Understanding open source versus closed source. What does that mean? You know, N8N versus make. Because N8N is a, if you have developer capabilities, um, that's more of an open source solution that you can kind of put some guardrails on all that kind of stuff. Really, really important. Um, and I do find a lot of times people don't think about that.

Speaker C: Any thoughts?

Speaker A: I think you kind of hit the nail on the head and when it comes to kind of managing all of this and you sometimes get to the Point of diminishing returns with AI, I think because you can have all the capability and resources to do something, but do you really need to do that thing? Like all the cameras in the world on phones right now, they have the ability to have like m. Multiple hundred megapixels on a camera. Then the AI will top it up and do uh, this and even more perfect picture. But then our eyes can't process it. So what's the point of doing that? Like, where are we going?

Speaker C: Why.

Speaker A: Why do we need to put all the time, resources and money into that? And I always give this example of you get to a point sometimes where if you're for example, applying for a job and you know, you have people that have their whole resume written by AI, then the job description was written by AI, then you, you submit your resume and then the resume is filtered through an AI bot to, to pick out the right resumes. And then, then you have like potentially your first interview process or your first, uh, interview set scheduled for you and that's sent out by an AI AI machine to schedule that interview. So who is actually being hired? Who is, who is the, who is the interviewee? And uh, you know, what's. It's like AI hiring AI. And is that something that the, the world needs?

Speaker B: Is it.

Speaker A: It does make things more efficient. But you have to then think about striking the balance of like supporting a need for, for the world or for society versus the need for commerce and business and making money versus something that will make the world better or something that we're just doing for the sake of, of doing. And it's definitely a tough balance. And of course, throughout, through different industries, it's, it's different, but it's, it's, you know, sometimes it's hard to weigh all of those things at the same time.

Speaker B: Yeah. If you're, if you're not thinking about the human interaction within it, right. You're completely taken away. Your proprietary capabilities, like if you just carbon copy and do exactly like what you're talking about, which is happening in so many cases. Right. Implement the capabilities, bring in the technology, but don't lose your human touch within that. Use it as a platform to increase your human touch.

Speaker C: Right.

Speaker B: And remember how important that human touch is, is because that's the thing that's going to separate you from people. Yeah. And that's the thing that we want, which is why things like verification is so important. We don't want to be talking to an AI all the time. Like that'd be like, do you want to sit on your computer all day? Hell no. That's why people still go to the store. But and so that that human connection and using AI to scale that um, is. That's where the winners are going to play. Because human intervention is still always going to be super important and we cannot get away from that. And if we think that we just want to be. We do not want to build a world of just robots. That's not the goal. The goal here is to help scale, help get us there, make our lives easier in doing it, um, but allow us to bring more human element to what it is we're doing, not less.

Speaker D: Right.

Speaker B: If that makes sense.

Speaker C: Yeah, it does. And it's interesting that you brought out some thoughts on scale because on a different perspective, like Philippe, for you, in a larger media organization wherein you might want to implement AI at scale, I would assume that a lot of change management has to happen.

Speaker B: Right?

Speaker D: Well that. But also I would say to your point, the talent, the talent to like it's a different way of thinking. Um, I, when you talked about what you just talked about, like a couple of things came to my mind. Um, one is like, you know, talents. You have to start thinking about this differently and how do you approach it? Um, I think you also your point around you said 85% of companies are like going to try implementations. Implementation. Yeah. Um, I think this is, this is an interesting take because so a couple of weeks ago had the opportunity to uh, I was at an event and Terence, uh, Kawaja, who is the creator of the Lumascape, had this amazing presentation on, on AI and uh, I think it was one of the best presentation I've seen in a very, very long time. I was fantastic. He basically separate, separated the, the AI, the future of AI companies and merger and acquisition into three different buckets. There was a first bucket that was like, you know, big media companies trying to invest into the AI ecosystem. There was a second bucket of um, startups that are, that are being built on, on the promise of AI and what AI can do. And it was a third bucket that he called posers.

Speaker C: Posers.

Speaker D: Posers. And those were companies that like have a. They say they do AI but in theory that it's just like machine learning, machine learning algorithm or whatever it is. Um, and I think this is the piece that everything a company has to be very careful with is like bring something to the game, don't just like, you know, use the term AI, um, to, to, to create a buzz, uh, try to bring something to a table that is Unique and that actually push, push the stuff forward for us at um, at, at ah, Chorus. For example, one of the areas like that, um, we've been debating and, and I read a lot of things around that is, um, if you really want some of those agents to give you the proper information back, you also need to feed them the information in, in in advance. And so one of the areas like for example having Global News, which is a very trusted news publication in the Canadian markets, right, that the hope at one point is also that some of that trusted news information gets somehow into those agents and gets sometimes gets somehow into those, those AI platforms. Because the one thing that we haven't talked about yet, uh, AI is also going to disrupt everything we do as a media company, right? Like, like we rely a lot on search and search is being replaced by AI. So how do we feed in our information into those AI systems so that when somebody types in a prompt about something, they actually get the proper feedback and they find us.

Speaker C: Right.

Speaker D: Uh, so for us that's also something that we really started thinking about. It's hard to think about when it's a very nascent ecosystem. Who do you work with? Who do you partner with? How do you feed this stuff in? Like where, where do you start? Where does it end? Like, how, how is it monetized? Like, you know, like we, we know how it took us years to figure out how somebody going into a search platform or like a search, Google search goes in and search for something and then ends up on global news. We don't know, we don't know yet how that's going to work with AI. Like, we don't know, we don't know how we're going to be featured. We don't know how those users are going to come to us. So that's, that's a whole area that we're trying to figure out as a media company and again, partnering with the right people along the chain to understand how you exist in that ecosystem. And then to your point, once you exist in that ecosystem, how do you bring that stuff back up? How do you bring it back up to clients, advertisers, partners, people that we work with to kind of close the loop. So it's, it's, it's, it's still very early in that stage, but those are kind of the steps that we have to go through and, and partner with the right people, partner with the right companies and the right people to make it work.

Speaker B: There's a, there's a really good podcast by the CMO of HubSpot called Marketing against the Grain. They showed Google Search, it's not going down. It's. It's actually continuing.

Speaker C: You have AI mode.

Speaker B: Uh, well, no, but Google Search, they're there. It's increasing.

Speaker C: Right.

Speaker B: However, the conversion rates are reducing. So conversions now, organic conversions, so we're talking organic, um, from. From SEO, have dropped substantially.

Speaker C: Oh.

Speaker B: Whereby only 1 to 2% of the traffic is actually coming from AI. But something like 50 to 60% of the conversions.

Speaker A: Oh, m. Wow.

Speaker B: Are coming from AI. So it's way less traffic, but the conversion rates are insanely high in comparison. Um, so you've got some of these platforms, like, Profound, and some of these other guys that are doing the AEO stuff. It's really interesting to see where that's where that's going to go, because that, that. But brand. Right. Like, kind of bringing this back to the advertising game. This is where we're pushing brands really hard and clients really hard. That brand is becoming even more important. The vanity metrics are going to start going away. And if you want to actually be asked, if you want to get in front of it, you need to be asked for. I believe video is going to become such a massive player in brand, so that when you're talking to your AI, you are specifically asking for the. The brand, not just the product. Because if it's the product, then it's going to be one AI talking to another. But if you specifically ask for Coke or you specifically ask for, you know, whatever the brand is, and that's going to come from understanding how to build organic content. That's going to come from how do you actually, you know, build really good creative, really good storylines, all that kind of stuff, and use those channels to drive that brand, drive that upper funnel, because in the end, that's probably what's going to matter most.

Speaker C: Yeah.

Speaker D: Yeah.

Speaker C: Talking about impact. And it's great that we're talking about this at that level, because I feel like Impact is something that we need to have a focus on. And going back to you, because we see that AI is transforming even how we consume media and how media is selected for us. I think that's one aspect that we need to focus on. Philippe, what are you seeing on that front, at least on. On the car side?

Speaker D: Well, so I think what we hear is, um, a lot of the. A lot of the work that we've been doing, um, so far has mostly been around how do we work with the partners in terms of, like, what do they want from us? What is the information that you would like us to feed into your different AI systems.

Speaker C: Right.

Speaker D: Um, that whole point I made earlier about the speed, that is usually the number one thing that everybody's asking for is like if I want to buy uh, an ad on course, how quickly can I go from I have a brief to I want to book something to what is the outcome? And that whole chain, if you think about it, from, I mean we've all been in a digital advertising business for a very long time. We all know how manual that can be between the insertion order responding to the brief, insertion order responding to the brief, and then the brief gets uh, dissected and then you respond back and then there's a negotiation process and then we talk about the rate and then we go back and forth and then we don't know if it's going to be programmatic or direct or whatever. And then if you start doing an integration, it even becomes more complicated. By the time you have this book, you have to deliver the campaign. And then usually there is information that you can pass along the pipes to give you real time information. But there's also scenarios where you have to create a post report and the post report becomes very complicated because you have different ad servers, you have different systems, you have all the different pieces and it all needs to come together. And we haven't even talked about the fact that data sits in the middle of all of this and we're talking about audiences, right? So I think what we hear a lot from our clients is that how do we go from that initial brief all the way down to audiences very quickly? Like if I have a brief, I'm trying to reach those, those very specific audiences. And to your point, I think this is where you was, this is what you were saying, where you were saying that brand is going to be more that the, all those metrics that we've been looking at for the longest time, they become less um, important in a world where as an advertiser, as a client, as a marketer, you know that the audiences that you reached actually are interacting with you. They interacting in a certain way and you get that outcome almost immediately. And I think that's what we heard a lot. Now look, are we there? No. Like we're not even close to like having that whole chain figured out. But it's something that we are thinking about. Certain agencies have built their own AI, ah, ecosystem. I mean we all know wpp, like they've been very, um, very open about um, how they look at the future. And this is one of the partners that we've started having discussion with about how do we, how do we feed our information there, how do we work with them, um, how do we get all the different pieces together. And that's, that's, that's really step one for us.

Speaker B: Um, you know, we're looking at systems like Airtable, um, you know, HubSpot, and actually starting to create systems where we can actually go from proposal to contract automated and that system. So what we're doing is we're building basically the framework right now of a system where we can build in a, uh, way to create a proposal through a form with multiple inputs and that gets captured. And then when that gets captured, you can renegotiate that. And all you have to do is change a couple things, but it sort of automatically changes the digital proposal. And then when that digital proposal gets signed, it's a, it's a digital signature that then triggers a workflow that creates an entire schedule based on when everything needs to be captured and creates triggers based on, okay, this needs to be in at this time, this needs to be in at this time. Um, and if it's not, there's another trigger that goes out, an email that goes to everybody. But the system is in place, templated based on exactly the parameters that it needs and understands. So we can start automating that process and get to these things quicker and even reconciliation down to invoicing and all that wonderful stuff. So that, that's where, that's one of the applications that we're looking at so that we can build that system, tie it to another table within, like an air table, and then because it's tied there, we can start bringing in things like N8N and build workflows and other agents and plug, start plugging those things in and also integrate that into things like tableau and so on and so forth and start building this kind of ecosystem that gets us to that point.

Speaker D: The step two is really in what I described, and you kind of like went that way, is how do you bring in the agents? Once you have those connectivities happening and you have figured out how the process and the workflow works, that's where you start bringing in the agents, because that's the extra step in, in the whole, in the whole process.

Speaker B: And, and just to give some advice on that, the only way to do that is to bring it into the cloud. If you don't have it in the cloud, then you won't be able to do that. But when you do that, it's really important that you Use systems in place so that you still own that data. So you don't want to just push data into the cloud. Yeah, of course you've got to have a system where you own it and you're normalizing it. That's what you're putting into the cloud and that's how you're creating those intersections. Intersections. So then you can start plugging these things in and start creating those automations. Really, really important. And that's where you know.

Speaker D: Yeah, we go, that's where you go.

Speaker B: It's a fine line. Right, so you know that and that's exactly uh, on our end we've, we've moved our dev team over to that specifically. It's like understand those things and focus all your attention on how we're going to plug all these things together, use the cloud in a privacy safe way, um, all that kind of stuff.

Speaker C: M that's a great segue though because I think my last question is that with all of the things that AI can do for us, all the opportunities that it can open, there's also that bit of responsibility that we need to have in place. So Seha, looking at you, um, how do you see ethical use of AI? Is that something that we need to be focused on and what does it look like?

Speaker B: For sure.

Speaker A: And I think to take it a, a little bit of a step back I, I think with the basic understanding that I think people find difficult to wrap their minds around or even if they can understand it, they don't keep it Top of mind is ultimately the data in all of this is the currency. Right? If you're using an AI tool and it's free, your, your data, you, you are the currency, your information, your data. Exactly. And that, that goes for the, for the consumer side but as, but also for like a publisher side or the, or the, or the manufacturer developer side for a company side. Because understanding that fundamental principle that the data is the currency eases up a lot of the issues when it comes to um, you know, getting clean data or getting data that is useful to develop, develop your business or generate proper, proper uh, learning and not non biased um, algorithms and things like that. But it also is very important for the privacy side because once you understand that that is the currency and people, if you're physical currency people are very hard to give away their dollars, their hard earned dollars. People are also probably pretty difficult to give away their, their currency, their personal information, their data as their currency. And if you take a look at um, the EU for example, the European Union, they're doing a great job with, with bringing on AI into the, into the forefront. Um, they've actually developed EU AI act, which is putting, um, privacy and responsible AI use at the forefront. And they're basically structuring it such that they're working with industry and stakeholders and they're putting that at the forefront and saying, you know what, we got to put this privacy first. And we're going to look at the individual industries and say, okay, well, what are the risks in this industry? What's the risk in the medical industry using AI? What's the risk in construction? What's the risk in education? And depending on the risk factors and the risk levels, they're assigning different levels of responsibility and obligation when it comes to incorporating AI. So medical industry health information is extremely important, extremely sensitive. They're going to have more obligations and more checks to check off and more hoops to jump through to incorporate AI into the medical industry and doctors and dentists versus somebody that is, I don't know, I'm using it to check NBA stats, for example. Um, so they're doing that kind of thing. It's kind of like a skill which I think is a brilliant idea. They're also incorporating, you know, we always use data privacy impacts. We do assessments on those kinds of things across the board. Um, they're doing human rights impact assessments. They're doing fundamental human rights impact assessments. They're doing, um, things like, um, uh, assessments for, for how the use of AI could impact the human, the actual individual person. And they're making those mandatory across the board for any corporation, any organization of a large size that wants to incorporate AI. Um, and they're, they're, they're doing amazing things to kind of bring the human element, as you keep saying, to the forefront and to make sure that we're, we're protected from, from that, from that angle. And um, you know, the final thing I'll say on that is I think it's important to, to kind of bridge that, that governance or the, you know, governmental policy and regulation with industry and industry stakeholders because it's a lot easier to make people follow a certain rule or certain law when they're involved in developing it. Um, and if you're taking into consideration what their pain points are, you can make sure that you handle those pain points, but then also keep the human element at the forefront. And that develops things like transparency in coding, transparency in how algorithms are learning. Um, letting people understand how those things happen from the back end creates trust from the, from the consumers creates trust from, from. From the industry and, um, and things like, um, you know, making sure that, you know, we've always heard that privacy by design, quote, but AI by design is a huge thing now. Um, it's, it's incorporating the idea from the ground level that AI is, has huge capabilities and can be a very useful tool, but can also be a very dangerous tool if not used and implemented correctly. And then, you know, you take away things like the algorithmic bias that can come just from, you know, just from coding and just from, um, you know, anybody putting in their own, you know, version of their life into how they code an algorithm. And, you know, it's not necessarily intentional, but it can create a bias in how outcomes come out and things like that. So, um, those would be probably the three or four different things that I think we should start implementing here on the North American side, um, because the European side is quite ahead of us on that.

Speaker B: I think this is where the use of synthetic data can really kind of come in. Once, uh, you get to a point where you've got the ability for an AI to run every scenario that requires compute, right? So like size and scalable compute. But synthetic data is potentially a way to help protect privacy as well. Because. Because if you just need a very small data set of data, human data, to then run every potential scenario through an algorithm that can actually use that and build synthetic data from it, then you only need very small pieces to plot out things like, um, precision or predictive algorithms. Um, and in doing so, you can also protect privacy in that way. Right? So the future is also, you know, how do you incentivize the audience, uh, on that front as well? Um, so there's, you know, like there's

Speaker C: a value for them as well.

Speaker B: That's the thing value exchange is. It's such an interesting thing. I mean, when we first started in the industry, that's all we did was value exchange and in mobile was inside of mobile opportunities. What are you giving this person to get them kind of thing. Um, the gaming, whole gaming, uh, system, mobile gaming was built off that watch an ad, get a credit, play a game, right? So I think there's a future where that value exchange can become a bigger part of the ecosystem as well. Um, but yeah, how those play together is interesting. Uh, I think maybe that's going to be. That's going to give us ways to enhance that privacy side.

Speaker C: Philippe.

Speaker D: The only thing I would add is from a legal perspective, it just feels like, I mean, this is what we see at chorus. I think we want to be as cautious as possible. Um, I think you said it, you know, you could get into very dangerous territory in certain areas if it's not properly controlled, um, and uh, structured and governed. Um, I think the piece also that, that we, I think we all have a responsibility as uh, um, industry, I guess to kind of figure out exactly what the guidelines are. Um, I think it's going to take some time. I'm not sure that like the talent really exists to really understand the full scope of what we're trying to do here. So I think we're going to learn. I mean look, we've been in this industry, all of us have been in the industry for quite some time and like 20 years ago when digital advertising started, there were also like quite some dangerous areas that everybody was trying to kind of figure out. So I think it's very early and I think it's very important for everybody to kind of like know what the guidelines are and make sure that you stay within the proper framework. And I think that's what we try to do, um, at chorus is like really making sure that, you know, whatever we do is. We know that what we do is like privacy friendly and we are within the guidelines. And I think I like what you said also about the EU being very uh, looking at this, uh, um, from uh, a bright. From a. From a very large scope in terms of what it means. And I think that's what we should be doing also as Canada, just having some guidelines and guardrails and making sure that the way we approach this is safe.

Speaker C: Yep. And certainly for us at IB Canada, we're keeping a close watch on this and uh, we want to be sure to help the industry as we all move towards, you know, use of AI in an ethical, uh, in a responsible way. Thank you very much for joining us today. Thank you and for our listeners for the business of digital podcast. Stay tuned for more conversations on how innovation is changing the face of media in Canada. Thank you. Thanks everyone.