Unlock campaign performance with AI

Uh, he's the client president at WPP Media Singapore. Hi, Carrie. And last but not least, we have Justin, and he is the director of China Partnership, APEC at Mastercard. So each of them brings a unique perspective on how AI is reshaping the campaign strategy and execution, not just to improve efficiency, but to truly elevate the performances. So get ready for an insightful and forward-looking conversation. But before we dive deeper, let's zoom out a moment

to look at the broader trajectory of AI. So, I often tell my friend that we are living through one of the most transformative moment in the history of AI. Think about this. So what begins as the rule-based. The system in the early days has evolved into neural networks, then deep learning breakthroughs. And today, we have the era of the generative AI, a large language model. And we have seen that AI has surpassed the human performances in many technical tasks today. So what makes, what makes this

ship so critical? It's, it's not just the technology itself, it's how we are using it. But move beyond just automating the isolated task. Today, AI is starting to orchestrate the entire workflow. From research from research and creative to the media planning, execution, and of course, the measurement. So what really that means for us in the media industry. So, it means that now we can create faster, smarter,

and ultimately deliver the bigger impact. So at Mediacorp, that's exactly what we are focusing on, driving the meaningful AI adoption across the whole chain. And in today's session, you will hear from our panelists about how their organizations approaching it. So I'm delighted to pass it over to my, to our first speaker, Lin, who will share how Kanta are thinking about growth in the evolving AI landscape. Lin, over to you. Thank you,

Canta is one of the world's leading brand and marketing data insights and consulting companies. We partner with more than 96 out of the top 100 biggest advertisers around the world, and we have presence across 90 markets and a global team of around 20,000 people. We help organizations understand how people think, feel, and act both globally as well as locally here in Singapore.

So thank you for having us again today and I'm here to talk about how AI infuses the brand market and marketing data and analytics that Canta provides for our clients and I hope that you'll take something useful for your organizations today. And when we ran a study as Cantor with uh global marketers around the world, uh what we found is that marketing is one of the most optimistic functions within a company when it comes to their embracing of Gen AI.

On the left hand side, the results of uh optimism scores from marketers is almost 9 which far surpasses any other functions within a business. And what most marketers are telling us is marketing will be transformed by Gen AI in the form of boosting human skills, making processes smoother, and therefore, it is critical for us as marketers to be early adopters of Gen AI. But make no um haste in this. There's still a lot much work that we need to do to unlock

true value. When we flip the page over to the right hand side, we can see that when it comes to how marketers are feeling the current impact of GenAI is on market. Within their organizations, it really still is early stage. A score of about 5.3 was given by marketers on how they feel about the potential of uh Gen AI, uh,

within marketing, but companies still limiting the existing usage. And when it comes to internal readiness, this is where it gets even more bleak, uh, a rating of 4.9 and what marketers are reflecting is there's a lack of training, a lack of, uh, tools that are really, uh, being given to for them to deploy. And therefore, it's imperative that we actually invest much more

in education and investment. So take this slide and show it back to your management on how the state of marketers and our optimism is really embracing, uh, our adoption of technology. But the limitations are really in the rest of the business. But have no fear because that's what we're really here to address today. Now, from Canter's point of view on the next slide, what we're seeing is really the limiting factor now for Gen AI and AI in general in marketing.

Is a lot of the times the use cases have been focused so far on replacing task, right? How do I get a um a meeting summarized more quickly? How do I actually write a copy much more faster? But the real opportunity we we believe with AI isn't embedding it in a way that transforms your end to

end processes. Now, what do we mean by processes? Every single day we as marketers go through different kinds of processes to get to an end outcome which could be a campaign or an entire uh year uh uh marketing strategy. The 3 main processes are respectively planning and strategy, which is really around how you're taking market data, consumer data and translating that into insights to go into your overall planning and strategy.

The middle part is integrated marketing communications. And that's what has happened when uh you take AI to uh create your end to end communications, uh, approach, uh, and going really from the early big idea to the creative and all the way through to how it gets deployed in market via media and being measured. And finally, on the right hand side, the last process is on innovation, which is a core role that marketing

contributes to as well. How do you actually innovate and develop new products off the back of the marketing, consumer insights and data that are being generated today. So over the next 3 slides, we will deep dive a little bit into this, and I'll bring some real examples to bring this to life. Now let's talk about planning and strategy and what uh AI opportunities there are for us to improve the way that we plan and strategize today.

In the middle part of this slide, we see the breakdown step by step processes that go into planning and strategy. They're starting off with your landscape assessment, which is really understanding what the macro, uh, market and customer uh landscape looks like all the way through to pinpointing the issues and opportunities, setting your objectives, translating that into marketing jobs to be done, then planning your activities, and then going

to uh measuring it. We think that there's several standout opportunities over here for marketers to embed AI even more. And one of the key areas is in issues and opportunities, for example, GAI can help you easily identify gaps by detecting patterns in customer needs and competitive weaknesses. And a good example of a company that has brought this to life is L'Oreal. They actually had a partnership with Nvidia uh to actually develop a gen.

to to augment their employees' ability to support growth opportunities, really forming a good partnership between some of the internal employees understanding human understanding of the customers with some of the uh AI synthesized landscape assessment and issues and opportunities that uh acts as a sparring partner to them.

Under objective setting, for example, you can use Gen AI to use and connect historical data and predictive analytics, set realistic data-backed objectives, and a good example of how BBVA has done this is they really use uh Gen AI as a sparring partner for them to create and refine some of the strategies going into the markets.

Well. And finally, on activity planning, this is one of the process steps where the use of AI will be a quick win because it can help map out activity plans from how you select your channels today to your messaging strategies to predict the optimal combinations for different kinds of segments. And what Heineken has done here is truly astounding.

They've used Gen AI to enhance their customer segmentation, which has led to more targeted marketing and significant marketing, uh, media savings, uh, in, in order to drive real commercial uh effectiveness. So that's planning and strategy. Let's turn the page over to the next process, which is really in the space of integrated marketing communication. I know many of us in the room today are actually responsible for the day to day, uh, uh,

execution of marketing. So this should be particularly interesting for you. Again, uh, in the middle part, we see the end to end process from how you develop the brief to the creative and testing it, creating the actual production of the creative itself through to media planning and measurement and optimization. And similarly, we see multiple spaces where actually Jay can

be deployed, deployed to actually make things more effective. Under creative testing, for example, Unilever has embraced digital training technology to create multiple copies of their uh uh creative assets and then implemented that AI to speed up and scale that creative evaluation. So they have actually tested 3 times more digital assets and 15% more TV assets as a result of the embracing of these technologies. And for media planning and campaign activation, a really nice example,

my personal favorite is from Catree India. They use Gen AI to create hyperlocal personalized ads featuring a very famous Bollywood actor, Shah Rukh Khan. And by analyzing the small business data, they actually use uh the uh customized ads to promote local stores with Shah Rukh Khan um by name, helping them to gain their visibility and recognition and collect, connect with customers during the festive season.

Finally, in measurement and optimization, TSB used AI to allow the ingestion of hyper granular daily data across both the digital and offline channels at speed. They refreshed their MMM model and we'll talk a little bit about what an MMM is, uh, later on. But they refresh this on a daily basis to enable real-time decision making. And I want you to pay particular attention to the last line over there. Uh, I know it's a little bit small, but it's in the TSB box.

They actually, uh, managed to increase their sales by 42% and reduce their media spend by 70%. Now, I think this is, uh, for most marketers in the room, and I used to be on the marketing side myself, uh, something so powerful that you can say to your end stakeholders, whether that's your, uh, executive or whether that is your, uh, uh, stakeholders internally, to be able to. Drive real, uh, commercial impact by improving marketing is something that is just, uh, becoming more and more essential in

today's businesses. And Gen AI, especially in the, uh, topic of measurement, is one of those things that can make a real differentiator in the way marketing shows up in the organizations. Now, finally, on the last one, which is around, uh, innovation. We see again a few opportunities over here. I'll just spotlight one in the interest of time. Our spotlight opportunity identification, which is really where marketing shines, uh, because of our

closeness to the consumer. Underportunity identification, a case study from Hellion was how they developed their Gen AI tool called Ask Halion, which is used as again, aspiring partner for innovation through the stages of articulating consumer insights through the way to interrogating the signs inside the products as a consumer health firm. And this really helps them come up with superior concepts that they can activate and test with consumers.

So those were the top 3 processes that we went through. Now, let's turn the page over to understanding a little bit more about how AI can be really deployed in extra solutions itself. First thing I'll say is this is not new to us at Canta. We've been around for a long time and we've been building expertise in AI for

a similarly long time. Now, what many of you, uh, might share in our journey is we've gone from in the early 1980s from machine learning, which is really around how do you do things in a much, uh, faster way.

Uh, to all the way, uh, in the middle of the part, uh, of the screen, deep learning, which is, uh, embedding AI to really help us understand and harness the data even better, to where we are now in generative AI, which is helping clients to take all of the, uh, legacy data that they have with us in order for them to understand the use cases and applications to improve into the future as well. And that's what we're really excited for.

So on the next page, you'll see that the transformation journey that Kanta has been on is that we've then started to infuse AI into every single solution that we have.

Whether that is AI for innovation, as you heard me talk about earlier, to uh make uh concept evaluation much more faster and much more effective through to AI for creative and being able to test at a lot higher scale as well as use predictive analytics to try to understand how creatives would perform even before they uh hit the markets, uh, as well as AI for media, which is uh optimizing your media allocation.

Uh, and finally, for brand as well on the right-hand side, in order to predict your brand KPIs, whether that's uh top of mind awareness all the way through to consideration and purchase intent. This is so powerful because all of our data and all of our AI is running off the back of a model that has a massive database over here. Uh, we use a model, uh, that's powered by something called Link, which is the world's largest normative advertising database.

It consists of 250,000+ tests, 307 million human interactions that have been recorded in our particular database, and therefore, we're able to predict how real human behavior will change as a result of seeing ads. The last set I'll leave you with is our clients typically see a 30% increase in ROI when they improve and add creative quality of the back of this insights from average to best. 30%, that's phenomenal. So I'm turning to the last pages in my presentation.

I just wanted to spend a bit of time just, uh, uh, bringing to life three core, uh, solutions that might be relevant and interesting for folks on this call. Uh, the first is, uh, what we call Lift ROI, which is essentially marketing mix modeling. It helps you understand, uh, of all of the media spend or all of the marketing spend that you're putting into, uh, the market, how much of it is actually generating true business results.

I.e. sales or uh if your financial services, assets under management, or any business metric that really resonates with you. Not only that, it can also be linked to brand equity metrics and be able to predict what is the lift in awareness, uh, consideration and purchase intent when you tweak your marketing mix or increase your marketing spend. In the middle, Link AI is our solution for really understanding how creatives would be performing in market.

And be able to predict how people would be uh responding to ads and on the right hand side, uh I evaluating concepts for innovation uh will uh require something that's called uh concept evaluate as well. You ingest some marketing data and non-marketing data, like your macro factors into a model. Uh, through ingesting of this particular data, it can actually, uh, simulate what the short-term sales and brand equity would look like in order for you to understand what is the exact tweaks that you

should be making in general. And if you're not using MMM in your organizations today, you should be really thinking about how you can actually deploy something like that in order to make your marketing much more measurable. On the next pitch Uh, I just spoke a little bit about how, uh, marketing comms can be improved. Uh, here's a list of some of the metrics that can be actually optimized off

the back of this. Uh, what's important to note is it's not just on the left-hand side, the brand and creative metrics, but you can also optimize your behavioral metrics off the back of this. Now, finally, on the last two slides. I'd like to uh just show you a prediction of where Cantar is going with, uh, the AI uh future.

We believe that today, whilst the majority of use cases is in, uh, the effectiveness, uh, area, which is really just summarization, automated reporting in the future under the Toor bar chart, it will actually change to become a lot more, uh, pivoted. Towards edge cases of AI. And what we mean by edge cases are things like contextual prediction to forecast not just the effectiveness of, um, advertising based on images or in music, but also things like the characters dialogue and

the contextual factors as well. So these are applications that are not common today, and we expect AI to be able to accelerate this into the future as well. So with that, um, I've come to the end of my particular segment. I know I kind of breezed through it. Please feel free to use the chat to ask me any questions as I see some of you have already done. I'm gonna hand the time back to Tim. I'm also contactable via my email that you see on screen right now. Thanks,

the next speaker. So let's hear from Kry who will share how WPP is leveraging uh its intelligent marketing operating system. WPP open to drive the performance, automate workflows, and reimagine how media get planned and activated. And over to you, Kari. Yeah, thank you, um, Tim. And so, good morning, everyone. Honored to be here. I'm Carrie and I'm the client president for WPP Media Singapore. We are the media arm of the communications group, WPP. Um, as, uh, the speakers, uh, as,

as we mentioned earlier, AI is evolving. It is also evolving very fast, very quickly, literally changes almost every day. Uh, I think one important point Lin touched on. And she gave the lay of the land is that more work needs to be done. There's a lot more value to be unlocked with the, with the potential of AI. Uh, I will speak from the angle of a user or of an organization using AI and weaving AI into our processes.

Uh, as we deliver for our clients. Uh, so I'll share how we have geared up and how AI is starting to play a big part for our organization and how we work with clients. Just a bit about WPP Media. We are a media agency, as I said, and we manage over 50 billion. Media budgets for our clients and our clients, uh, some of the logos, I've listed below, we expands uh global clients, local clients, both public and private sector. So the aim is for us as an agency, as a business, is

continue to be a trusted advisor to our clients. Uh, as an organization, uh, just like yourselves, we are upscaling our entire workforce. AI is, uh, as I said, it's evolving on a daily basis, so there's a lot of work to be, uh. here. Uh, our advice to all companies and all the listeners out here is that we have to stay nimble and age out in order to thrive in this dynamic environment. If you go on to the next slide, um, I will touch a bit about how AI is being used today in our organization, um.

The first one is a personal productivity tool and um I'm sure many of you have many examples of how you are using AI today from a personal level. I mean, you're doing searches apart from your Google search and your Bing search or using LM search. And so on. You're using this to summarize documents to generate ideas and so on. So, the productivity to angle,

I think it's um it's easy to visualize. Uh, as a media planning agency in the context uh of media planning and activation and measurement, uh, we have leaned into AI in a couple of ways, uh, to discover new audiences. What this means is, uh, looking for new sources of growth or for client growth. Uh, for clients, businesses, uh, planning media campaigns, uh, that are turbocharged in terms of getting outputs quicker, so we become more efficient and, uh, Easily what can be done in 2 days when it

used to be 2 weeks. This doesn't mean that planners become redundant, but AI has allowed us to have richer conversations, more conversations and discussions with clients throughout the process. So yes, there's efficiency, but I think the important and salient point here is that we get to better outputs as well. Uh, from, from a media reporting standpoint, it's also improving by leaps and bounds. We don't spend a lot of time.

Now just providing observations, it's not just about compiling data, it's a lot more time spent on understanding uh the insights and I think importantly, what are the actions, um. Uh, that will come out of, of all these uh insights, right? So how do you optimize the campaigns and so on. What are some of our clients telling us? So, as, as a media agency, our clients are equally curious, they

are consumers themselves. They are, they're asking questions as basic as, you know, what, what is this, uh, what is the AI landscape about? They're asking about what the AI capabilities are for AI agency, meaning us, right? So as a media agency providing our media investment services to clients, what is it? How are we using AI and how are we staying relevant for their business. So we're being asked these questions

uh quite a lot. Uh, some clients are asking us to connect how we do, uh, how we use AI to how they're getting tangible business outcomes, especially for clients that do a lot of business on .com. Uh, so those are very common questions and we continue to get these questions, uh, on, on a very regular basis. We're turbocharging ourselves through this, we call it an operating system,

and this operating system we have named WPT Open. Uh, we have committed to investing over 300 million to building cutting edge capabilities within this OS. It combines some of this development, uh, our own development, but it also um uh combines this with the AI capabilities of uh our key partners including Google, Amazon, Microsoft, Meta, and, and even TikTok. I hope you enjoyed that and you got the gist of what WPP Open is allowing our teams uh to do across WPP and WPP Media.

The what is really some of the platforms, some of the some of the uh interfaces that you see in the video, um, it's a unifying um area where we start doing our work from. It's the first thing that we open when we get to work instead of uh emails or apart from emails. The how is the nuts and bolts of it, but I, I think the how is also how the organization is bringing along all the employees and the workforce and

how we're enabling them to do the work. So, uh, an important factor to consider here um is that change management, right? So AI is not going to be a silver bullet for your business just because you implement something. You have to think of um some of the processes. I think Lin mentioned that a lot of um It is not just about productivity and automating some repetitive tasks. A lot of the, um, a lot of thinking is to go into how to bring everyone along this journey.

And this is especially tough because it is evolving so fast. The, the why, I don't think I need to kind of labor the point. It is, it is, it is a necessity, like it is going to be the differentiator, if not the silver bullet for many organizations. Uh, so for the BPP open on the next slide, uh, I'm not going to label the point here again. There are different, um,

Uh, what we call command center. This is the command center of marketing operations that we call it studios and uh different parts of our business use the different studios in different combinations. In the next slide, you'll see how WPB Media, as the media agency, the studio that we use the most often is Open Media Studio. Uh, you can move on to the next slide. Yeah. So this is the studio that we use most often.

Uh, alongside the Creative Studio. And um we approach as a media arm of the DPP, we approachedri clients and media problems through a process called DPAM and in each of the phases of Discover plan activate and measure, we have different tools that are sit under these processes. Uh, let's look at some of the examples. I think many of you are using uh AI in a personal productivity, um.

Uh, capacity for discovery and so on. Uh, one of my favorite use cases, so I thought I want to share this, is, um, the ability to turn long PDF documents into podcasts that you can then consume as you're on the go, when you're on when you're commuting. I think this is a really good personal productivity tool and uh we have this weaved into um WPT Open. So it's something that a lot of colleagues use um at a personal level.

In the next, uh, example, this sits in the planning phase within DPAM and um This, what takes what took us a week um to do can be achieved in a matter of hours, but let's call it a day because I don't want to give the impression that the AI spits out work and then this is what we pass on to the clients, right? Um, at the bottom left of the illustration.

Uh, what happens here is, this is an example of um a school, an education center giving us a brief on driving awareness for new postgraduate courses in the next school year. So in the past, this would have taken quite a while to compile across different data sets, um, different research tools that we subscribe to. Uh, right now, WPP Open and AI allows us to spit this out in literally minutes. The outputs um would be a draft of the touch points uh for

this um particular prompt, right? The prompt is how, how do we, uh, what, what should we, what are the consumer touch points that we should consider uh as we roll out and drive awareness for this course, right? So that The draft is given to us in terms of what are the digital channels you can possibly use, what are the physical channels you can, you can use. It even breaks it out into different stages from, from a teaser phase or awareness phase to consideration and then to a

sign-up phase. Um, this note, this is not um taking from open internet, we're pulling from a syndicated third party data. Um, the next speaker, Justin, will speak about this a bit. Um, but it is also um pulling from proprietary data from WPP itself. So it's important to note again that, you know, this doesn't mean that we just, you know, throw this across the the fences to the clients. Uh, what we are achieving here is, um,

Spending more time, validating, checking. More importantly, we are idating how to bring to life the campaign. We're working with creative agencies and clients uh in terms of messaging, better messaging, more accurate, more personalized messaging, better formats, uh, formats that will achieve better results and eventually better business. Uh, so Netnet, we are getting to uh better outputs faster, right?

Um, the use case that you have, um, here as well is the part of planning as well, and it's part of, uh, it's where we go deeper now to consider channel efficiencies, uh, different types of buying um um mechanisms to achieve the best outputs possible. So we are doing all this and getting campaign benchmarks from the category, from specific campaign data a lot easier. So I mean, what is the big deal in all of this? I guess, um.

If you think about it, the, what in the past, what we have, I mean, I, in fact, some of you are probably still doing this, you know, opening up different folders in your computer, finding different Excel, different pivot tables, putting it onto a a big screen and then kind of combining and getting. The core relations, so I'm, I'm sure a lot of us are still doing this today. Uh, from, from the point of view of a media planner, what we are doing now with AI is simplifying and, and solving uh

for time, right? It is quite a big deal. Um, the, the analysis. I'll give to you is that instead of using a mobile phone, try going, put a coin into a pay phone and, and, and make phone calls that you will never go back to those days, right? Your mobile phone is with you all the time. So this is the reality today. In the next couple of years, new media planners joining the industry will not know that this wasn't the norm before. And the last um Use case I'll share is this one called uh agents.

uh Lin touched on uh generative AI. There's another branch of AI called Agentic AI, you might have heard before. This is used quite a lot in within WPP Open where we are creating agents, um, and agents are like the approximate of a real person. So we, we use this to create potential um audience with um. Target audience, potential um uh experts who can weigh in on, uh, say your creatives or uh being part of a focus group. In this case, we have created um

The approximate of an intern, right? So we call this um the intern agent and we have built many interns for different clients. In this particular use case, um, this agent uh or this intern, uh, we have fed it a lot of the campaign data of um the uh a specific client. So all the. campaign reports and all that, we have trained this agent up such that we, we can ask this uh agent a number of questions. In the past, we would have to kind of again look into different folders for uh

data and all that. Right now, it's just a prompt like send me all the post campaign data um from this period. And the agent will respond, actually, this data, you know, it doesn't sit within this period. Would you like to

look at it from another period? So it becomes a very natural way of finding data and I did not share the full um screenshot because of uh client confidentiality, but um what this conversation continued to was bringing up different data and then uh different prompts to bring bring us closer to what exactly uh clients are looking for. So this is an example of uh agentic AI and it's getting a lot more sophisticated. There are a lot more interesting use cases. If you can share the next video.

Where AI innovation meets the art of storytelling, creating a new way to tell stories through our nation's languages. When we first embarked on the project, we wanted to really find an opportunity to explore the possibilities of AI because art, music, storytelling is such a big part of our culture. First, we went through a lot of folk tales to finally settle on. One that was appropriate, kids friendly, and we also settled

on the story of Pulau Ubin. We invited eager parents and grandparents to try out our tech while assuring them that the data would be stored and handled securely. Making sure that they feel comfortable with using our AI tools was a big success to me that they actually trust us in handling their data. With their full consent, they were ready to experience narrative. Over 40 AI tools were experimented with before we shortlisted 6 for our AI experience. We captured facial.

Thank you, Kerry. I think that's was really a powerful look into like what's possible when AI is, isn't just a tool, but it's actually becomes the operating system. And I think AI may be, you know, the engine, but we still definitely need the fuel, right? In this case,

I think data will be the field. So that's, you know, uh, let's welcome our next speaker, Justin, uh, who will share how Mastercard is using their data and AI to unlock a deeper audience inside, enabling personalize the scale and power more intelligent marketing outcomes. Justin, over to you. Thanks and um good morning. I hope you can hear me clearly. I'm just starting from Melbourne today and very happy to be here. I just a quick introduction about myself.

I lead channel partnerships and alliances for Mastercard in the APEC region, and we work very closely with our ecosystem partners like WPP Media as well as Media Corp. and just prior to Mastercard, I spent close to 14 years across the different big tech platforms from Google to Mata to TikTok. So really have had the opportunity to have a front row seat on the development of AI over the last decade and a half. So we've heard a lot about

the exciting possibilities of AI from Kerry and Lin. Um, one thing is certain in the age of AI. Data is more important than ever, right? As I think one of the panelists answered this, it is one of the challenges today is access to data, but more importantly, access to high quality data and as they always say

garbage in garbage out. Um, so hopefully today we can share a little bit more about what Mastercard is doing in this space to unlock value with our transaction insights and how we play a role in supporting media campaigns and delivering value across the customer journey. So I'm sure all of us, or most of us in this room, you have heard of Mastercard. You know Mastercard is a payments and credit card company, but not many of us might know that over the last 50 years,

we have evolved to become a data and technology leader globally. Uh, as you can see on this slide, our superpower lies in the depth and the breadth of data that we see across the Mastercard payment network. And on an annual basis, we see more than 200 billion purchase transaction points sourced from more than 3 billion cards in circulation today. Across more than 150 merchants, million merchants, but importantly across

a few 100 industry categories. So again, just to emphasize the depth and the breadth of the data that we have, and we leverage all of this aggregated and anonymized transaction insights and expertise to serve more than 4000 customers across 120 countries today. So very big numbers that I'm throwing out there, but I think as you can imagine, this really speaks to the trust and the quality of the data uh that Mastercard is holding today. Net Mastercard data is not just an asset, it is

the core of our network. Every transaction, every interaction, every insight is powered by the data flowing through our systems. And as I mentioned earlier on in the age of AI, quality of data is most important, and AI can process information and unprecedented skill and speed, but it's only as good as the data that it is set. So if your data is flawed, if your data is incomplete, or if your data is biased, then your AI outputs will be also. So as I mentioned, garbage in garbage out.

And that's why Mastercard has invested so heavily in building this trusted, anonymized and high-quality data ecosystem that spends 200 billion purchase transactions. This is because the depth and the breadth of data allows us to drive innovation at scale, helping our partners, advertisers and brands like you. Unlock performance across marketing, payments, as well as customer engagement.

So whether it's powering hyper personalized campaigns or enabling real-time decisioning as what Carrie had mentioned across the life cycle of media, our AI capabilities are built on the foundation of clean, structured, and actionable data. So as you think about your own AI journey, just remember that data integrity isn't just a technical concern, it is a strategic imperative. In today's fragmented media landscape, one fun fact I learned today is. On average, all of us are exposed to more than

20 plus different media channels. So that's a lot of media channels that we're all being exposed to. The marketers face a very growing challenge as to how do you reach the right audience with the right message at the right time. And also proof that it works in driving business outcomes.

And this is where transaction data becomes a game changer because unlike previously proxy signals like clicks or views or interest behavior, transaction data reflects real world consumer behavior because if you are a top spender in men's apparel, you are likely going to be in the market to buy men's apparel. So what people actually buy. When, where, and how these are where transaction insights really come in.

So by leveraging into anonymized and aggregated transaction data, advertisers can make media more intelligent. Right, where you can use real world market and audience intelligence to guide media planning as well as targeting. You can make media more personal by delivering hyper personalized experiences based on actual spending patterns and preferences of the individual.

You can make me more measurable, where you can link campaigns to tangible outcomes like incremental sales, and no longer just engagement metrics or vanity metrics like impressions of views. So for example, instead of targeting urban millennials interested in travel, We can identify micro geographies where there are high frequency standards in travel categories where they are actually transacting and actively transacting, and you can then tailor those messages accordingly.

So this level of precision helps advertisers move beyond assumptions and generalizations, moving to data back decisions that can drive performance. And we look at how this translates into real world media and advertising workflows with data.

So in a world where consumers are increasingly engaging across multiple channels, online, in-store, mobile, social, we all need a way to connect the dots across the full customer journey, and this is where transaction data becomes a powerful enabler, because it provides you with a direct lens into what people are actually buying, not just what they're browsing or clicking on. So if you look at number one here.

Mastercard's data can help advertisers and brands like you leverage real world market and consumer insights to a more effective media planning. So, for example, we have transaction data across a few 100 industry categories. Across the time series. So as a brand, if you want to understand when does retail stand in Kits apparel pick up leading into mega sale day, such as 99 or Christmas in Singapore.

This data will help you to understand that and help you plan when you should start your media plan or the media cycle. But we can also look at benchmarking of year on year growth in the kits apparel industry so that you can better understand how is your business performing against your competition. We can also help you understand online versus offline spending, so that helps you understand media budget allocation to offline and online media channels.

And more importantly, we can even go down to a brand level to help you understand market share, like pre and post a media campaign. Are you actually gaining or decreasing in market share against your competitors? As you can imagine, all of this data becomes very powerful in the world of agentic AI as Carrie mentioned, because as you query an agent, all of these insights will really power that model.

2nd, #2 here is we can help to enhance audience targeting by reaching high value and high-end customers, right, where you can identify where all of the top spenders or frequent spenders on hotel bookings or travel in Singapore, and you can then serve them very targeted and personalized creative messaging. Yeah. And the last piece of number 5 is where we can then look at how do we measure close

loop attribution. So MasterCard owns a technology called Tess and learn for media measurement, where we have the ability to understand what was the incremental sales slip that was driven off an omni Media channel campaign and also understand those business drivers.

And all of this becomes very valuable data that we can inject into AI models across the entire media planning flywheel to help automate and democratize research, audience planning, as well as attribution to make media a lot more actionable and dependable. So hopefully this gives you a very quick sense of where our data comes in.

Uh, and then on the next slide, I just want to walk through a very simple use case of how we have partnered with a leading airline in this region to apply AI in a very practical and high impact way. So the problem statement that was brought to us at Mastercard was, how can we help this airline partner optimize promotional mechanics and customer engagement strategies using data and AI.

So working together with them, we co-created past specific AI agencies what Kerry had mentioned, these are like your interns, your internal agents that can leverage MasterCard's data insights, platform capabilities,

and our deep analytic expertise. And so these AI agents help to power and scale in two areas, knowledge management, as well as insights generation for the airlines in-house teams, and it helps them answer key questions like What are the key insights from previous initiatives to help me guide customer engagement strategy? What discount levels did I offer previously that were able to drive the highest revenue uplift for my business or which SKU or which flight route or brands overperformed during

the past campaigns, right? And lastly, what promotion periods you did the best results. In addition, we also supported the area of customer segmentation and personalized messaging to identify high-value segments based on historical behavior and the agents, in this case, the AI agents that help to tailor email messages, in-app messages by dynamically using customer level metrics like redemption rates or booking patterns.

So the result is a more intelligent and responsive marketing engine, one that is grounded in real transaction data and is capable of adapting to customer behavior in real time. And this really goes beyond just traditional AI, but how AI is in action today with Gen AI as well as predictive AI. So I put on to the next slide. And this is just another tangible example of how Mastercard

is helping our clients bring AI use cases to life. First, we provide our data and insights in multi-formats that are optimized for rack. In this case it's retrieval augmented generation to support large language models as well as AI agents. Second, we share instruction templates and best practices, including prom libraries, sample Q&;A, and tested use cases. And third, we have our in-house digital labs team that helps provide consulting.

To design a prototype and build these custom AI solutions. So the goal is to help our clients like you move beyond experimentation, operationalization, and building AI into your workflow in a way that's measurable, secure, and future. So these are just the 5 e takeaways. In summary, generative AI is really redefining consumer engagement today. We believe that transaction data is a key strategic asset and it is a competitive differentiator to brands and marketers.

With the use of AI today, personalization is now truly scalable and AI powered is also a lot more accountable to you as well as the decision makers and leaders in the company. So hopefully this gives you a a comprehensive or even a concise overview of where data plays as a key enabler across the entire world of AI. And with that, I will take a pause and pass the time back to Tim and Chai. All right, thank you so much, uh, Dustin. I think that's a great sharing.

So I think for the interest of time, um, let's just move to the next section, which is the Q&;A. I think some of the questions has been answered by Lin and Carrie in the uh in the, in the chat uh section. But I think I'll just pick up one of the common questions being asked like what are the tools that are being, uh, what are the generator tools that are being used and also what's the differentiator if one is, you know, use the same kind of tools, right? So I think.

Uh, actually, the, uh, Justin summarized it in his takeaways. I think one of the points that he made is actually, of course, data will be one of the key differentiators. I think if you see from all the three, panelists who was sharing, so what is the common is they started to use the tool, they started to actually building their own workflows. You can tell from, you know, uh Carris DPAM and also uh uh Justin also share their own master class, uh, you know, the over, uh, the,

the frameworks, right? I think. That's definitely something that common, but I think what's difference is how you are integrating the data and your own data, right? And then also basically fit into your own workflow because that will make a bigger differences. And many of those tools are model uh uh large language model, um, agnostic, meaning, whether you're using JGBT or use a different, you know, tool, actually, you know, it.

Differences, but when you are connecting that with your own data, with your own workflow, I think that's how you basically differentiate from other players, right? I think with that, I would like to thank, uh, you know, three of our distinguished guests, panelists, for giving uh the great sharing, and I think we learned a lot. And I also want to thank everyone, all the participants, all the audiences joining

us today in the morning. Thank you so much. And hopefully, we're looking uh forward to see you in the next session. Thank you, everyone.