- yesterday
Variety Strictly Business Live presented by Deloitte and AWS
Category
✨
PeopleTranscript
00:00Welcome. Thank you for joining us today. This is a first for Strictly Business Variety's
00:12weekly podcast featuring conversations with industry leaders about the business of media
00:17and entertainment and sports. Today we're going to be talking a lot about sports. We
00:22are at the beautiful Cannes Lion Festival of Creativity in the south of France, the
00:29French Riviera, and it is gorgeous. It is living up to the blue in the name of the Côte d'Azur,
00:35and we are just immersed in discussions about media, entertainment, fandom, reaching audiences,
00:44using new tech tools to unleash all kinds of creativity in ways that we can't even conceive
00:51of, but we will by the time we get to Cannes Lion next year. So really, really excited to be here.
00:57This is the element for the Strictly Business audience, and we're very excited. Today we have
01:02a terrific pairing here of two major companies that everybody has heard of, and what's really
01:09cool about this pairing is that both separately and individually, Deloitte and AWS are doing very
01:15cool, very innovative thinking, forward-looking activities together and separately. We have
01:22Michelle McGuire, Principal and Chief Commercial Officer of Converge by Deloitte, a very big new
01:29initiative that we were going to talk a lot about. And next to Michelle is Ruba Borno, Vice President
01:34and Global Specialist in Partnerships for AWS. Thank you both for making time. I know it's a packed
01:41schedule here. The place is hopping, so we really appreciate you giving us this time.
01:45Thanks for having us. It's so beautiful here. Great weather, so could not be better positioned
01:50to talk about sports and media and entertainment. Absolutely. Thank you, Michelle.
01:54Very inspiring space. So thank you, Cynthia. Thanks, Michelle.
01:56Again, thanks for making the time. So as I said, you guys are both doing very cool, innovative
02:02things separately and doing some work together in partnership. Let's start there. Let's start with
02:08Converge because that's new. And also every other conversation here involves sports. Everybody is
02:17talking about the growth and just the number of not only the big sports that we all know, but niche
02:23sports are really, really starting to become, you know, certainly the niches are getting bigger in
02:28many areas. And that is a huge opportunity for media, obviously, with sports and the importance of
02:34it all in that. It's just so much in the fabric of media and success for platforms these days.
02:40Why don't we start since it's relatively new? Michelle, why don't you tell us, just kind of
02:45give us that elevator pitch on what Converge is? Sure. Yeah. So a few years ago, Deloitte decided to
02:51start investing ahead of the curve in all things AI and sports. And so Converge was born with the kind
02:59of intent of, at the intersection of AI and industry. They were coming together to create
03:05differentiated, accelerated solutions for our clients. And so the firm has invested over a
03:11billion dollars at that intersection and gladly partnered with AWS to build out, you know,
03:17industrialized products that are available to clients to accelerate them to value, to outcomes,
03:24and to impact. Ruba, can you talk about sort of how AWS plugged into that and why you were the right
03:34platform? It sounds like from what we've been talking about, it sounds like you were kind of
03:38the only platform that could enable all the different, you know, all the different functions
03:43and applications that Deloitte was looking for. Yeah, our partnership with Deloitte goes a really long
03:48time. I mean, Deloitte and AWS have been partners for over a decade where we have been serving clients
03:54together, supporting them first on their cloud journeys. So initially at the inception of cloud
03:59services, Deloitte and AWS would go together to customers and help them migrate out of their data
04:05centers, which were energy inefficient and also didn't have the innovation and the number of services
04:10that AWS could offer. Today we have 240 plus services. And so with Deloitte, where they have a deep
04:17understanding of the industry, and so you take even sub-industries. So within sports, it could go into
04:22different types of sports and different audiences and sports. And then we've got the technology,
04:27underlying technology services that can support them in delivering some of those outcomes. And so
04:33we've been working on business outcomes and solutions with Deloitte for a really long time.
04:38I think that's where the partnership comes to life, is how do we deliver an outcome to a customer?
04:42And that's what's really exciting about the Converge platform is the outcome is a differentiated and
04:47unique fan experience. And it's in sports, but it can actually apply more broadly than that.
04:52Yeah. I think AWS is uniquely positioned given their, obviously the infrastructure associated
04:58with cloud, but also with their AI enabled tools. It allows us to tap into their subject matter
05:05expertise, their engineers, they support us from a product perspective. So for us, it's always,
05:10you know, easy to go with best in breed. And so partnering with AWS is always a great choice.
05:15Right. And I think, I mean, obviously, if you have, if Deloitte has put a billion dollars into this,
05:20you see massive opportunity on the horizon.
05:22I think so. And you mentioned it in your introduction, right? Like AI is everywhere
05:27here in Cannes, and you would imagine that a festival of creativity would be about, you know,
05:32human, you know, innovation. However, it's about how you partner with the tech and with the technology
05:39to transform the creative process for a different outcome. And it's no different with fans and sport,
05:46right? They expect a personalized experience. And that's what Converge does is it enables fans to
05:53identify, raise their hand, and then for us to basically understand what is their affinity?
05:59What is their propensity to engage, to buy a ticket? How do we, you know, understand your fandom score
06:06of, which is effectively lifetime value of those fans to both sports organizations, as well as to
06:13partners who, or brands who really are about the coming together of the league, the team,
06:19and, or the, you know, the city like LA 28, right, coming up soon. So I feel like, you know,
06:26there's a bit of a push towards all things fandom and the ability to measure it in a way that
06:33historically we could not.
06:34And just maybe Michelle building on that, because I do think with generative AI and
06:39personalization, partnerships are the future. It is the only way to provide that personalized
06:44experience.
06:45Because nobody has all the data and all of the ability that makes sense.
06:50There's just much more value out of bringing the data together and the insights from it. So
06:54actually a service that Deloitte uses is AWS Clean Rooms. And this is one where you can bring
07:00data from multiple parties and the raw, the underlying raw data is still secure and owned
07:07by the entity that brought that data, but you can draw insights from the collective data. And so that
07:12allows us to get insights, uh, across, uh, one of our customers is Coca-Cola, for example, and they're
07:18able to use it to provide aggregate insights, um, to their, uh, advertising team, their marketing teams,
07:24and then be able to provide a personalized experience to their customers. Another customer I would call
07:29out is, uh, the weather channel. They worked, um, with Latamy on a travel and hospitality customer
07:35of theirs. Um, and they're pulling data from a bunch of different sources and using clean rooms.
07:40They're, they care about the insight, not the underlying data. They care about the answer,
07:44which is, is this individual, a low, uh, frequency flyer or a high frequency flyer? What is
07:50their brand affinity? What are the brands that they are associated with? Do they travel by air or land
07:55or by sea? Um, and then as a result of that, of those attributes, they're able to actually get
08:01those insights. I think the number I saw was 98% faster. Um, and the entire process and the cost to
08:08run it is seven times more efficient, um, than what they were doing before. And so that's the power of
08:14partnerships is you get the answer faster. It's lower cost, but more importantly, you're able to
08:19provide that personalized experience to that end customer. And I think that for consumers, for
08:25customers, for advertisers, this is table stakes today. Um, and so the more data you have, the
08:31better insights you have, and you can be differentiated. Right. You are going light years
08:35beyond age, gender, general geographic. To your point, it doesn't matter whether that person who
08:42flies all the time, you know, lives in Cleveland or lives in Los Angeles. I guess some of it matters
08:48according to their airport, but the larger point is that you're able to just find those,
08:53those discrete pockets of potential audience, um, in, in a way that is just, it's fascinating.
09:00It's three dimensional chess versus when I started, I would get a fax. So this dates when I, when I
09:06started covering television, I would literally get a fax of the overnight ratings. And when I could see
09:12the grainy numbers, sometimes they'd blur together. I could see the grainy numbers and I'd look market by
09:17market. They did 50, you know, people, people in television know this 50 markets or 56 markets.
09:23It got to 56 markets. That was 70% of the country. And we scrutinize that and look for the patterns.
09:29But this is a, just incredibly writ large now with converge. One of the big, big use cases for you is,
09:35is really, really connecting sports teams, events, games, and sponsors, and just making them forgive me
09:43home runs every time. All the sports analogies are incoming. I'm really showing some discipline here.
09:49And what happened to your baseball hat? You had it on earlier. I think we're going to need that back
09:53for this many sports analogies. Um, it's a great point because if you are a sponsor of a team or a
10:00league, what you were interested in is that Venn diagram of the team, the league, and your own brand.
10:06And plus what converge then offers you is a, basically a data fabric of 260 million adult
10:15Americans. We know who you are. We know what you care about. And then we use our, um, propensity
10:21modeling, our IP, our understanding of sports and fandom to hand that off to sponsors so that they're
10:28then able to activate in a more personalized way to your point. Right? So it's about what player do you
10:35care about? What merchandise do you buy as a result? And how do I get you to kind of engage with the
10:42brand across the lifetime of fandom in a way that makes engagement feel more personal and like it's
10:51going to continue to engage me with both my team and the brand at that intersection.
10:57The metric is the amount of time. Well, the metric for the sponsor, obviously, but the metric that
11:01everybody's looking for is that time spent is what you say is that so we've actually evolved this idea
11:08of lifetime value into something called a fandom score. So it looks at all of the attributes associated
11:15with your engagement. So that might be in person, that might be online, that might be social, it might
11:20be your sentiment that you're putting out into the universe into the social universe. And then it applies
11:25some of your behaviors, whether that's purchase behavior, or it's browsing behavior, in a way that
11:32is, you know, frankly, a bit evolved than the traditional time spent viewing, which we both come
11:37from a TV background, I spend lots of years at Turner Broadcasting, like Nielsen used to be the metric,
11:43and it was yes or no, did your eyeballs watch this content? Yes or no. So now to think about we now have a
11:48metric that is 24 different data attributes associated with how much you love a brand in your
11:56fandom is, you know, light years ahead of where we were, you know, just a few years ago.
12:01Ruba, were there anything in helping Deloitte put Converge together? Were there any, was there any R&D,
12:07any innovations that that AWS did to empower what they wanted to do?
12:12Yeah, there's a wide variety of services, but maybe a couple that I'll dig into is Amazon
12:17personalization, which actually, you know, you can put in the data, whatever the customer is puts
12:22in the data, it then pull a data, there's data like user, their interactions, clicks, time spent
12:28somewhere, any purchasing behavior. So it's not just what you tell me you want to do, but what you
12:33actually do that we're now measuring. And that's, I think, what the personalization is all about,
12:38then taking that, and it actually selects the data to train a model that is based on the user's
12:44actual behavior. And then that model is used now to predict what the next user is going to do based
12:50on similar attributes. So that's just one example of what we're using. And then the other piece is
12:57Deloitte's been an amazing partner with us with Generative AI and using Amazon Bedrock,
13:02which is our managed service that allows the customer and the partner to utilize whatever large
13:08language model is fit for purpose for that use case. So whether the fit for purpose is analyzing image
13:14data or taking real-time speech and then converting it to text or getting a whole bunch of data points
13:20and putting out insights. So one example we were talking about earlier, Cynthia, was Formula One.
13:27We worked with Formula One on Track Plus. And in this solution, you've got, it's not like there's
13:34one ball in a stadium. Not that we don't like ball sports, but this is a sport with 20 drivers.
13:39They're going up to 230 miles per hour around a racetrack. And there are 1 million data points
13:46per car per second. Now imagine your former commentator self that's like reviewing that
13:52and trying to give a fan some useful information. Oh, and by the way, this isn't the only race they've
13:58raced in. There's historical data. They've switched the tires because it's wet or dry, or maybe they chose
14:04not to switch the tires for whatever reason. All of that information now can be provided in real
14:10time, contextualized historically and giving the why behind it. And now the experience for fans is
14:16much richer. And this is a very global audience, 500 million plus fans. And I think anyone that
14:23watches Drive to Survive like that took the viewership up. So very diverse group. And now you're
14:29able to tailor these insights to that audience. That is fascinating. And anybody that we'll see
14:35in another week or so, we'll see the F1 movie can highly recommend. Well done as an entire package,
14:41but you will learn. And as fans of Drive to Survive have seen, there is a lot of engineering and math
14:48and just pure F1 is, I think, you know, particularly in America, we're really starting to realize just
14:54what an intricate sport that is and how much goes into it and how much Lewis Hamilton and all those
15:01drivers, just what they're doing while they're in that tiny little claustrophobic car sweating out
15:08the G-forces. It's really fascinating. But the amount that is just that the drivers are having
15:13to process, I would imagine that there's also some benefits there too. And just even helping the
15:18drivers understand that they understand that tiny little infinitesimal amount of gas that's left in
15:24the car if they can go two feet more, two meters, I should say, further than the next person.
15:28Think about it from a fan experience, because some fans want the stats and the technical details.
15:33Some want like the history. And then there are some that like really want the drama. I mean,
15:37they want like, you know, they banged the steering wheel or they were really upset or threw their
15:43helmets or whatever. You know, that was actually something that we learned working with Bundesliga,
15:48which is the German football association, professional football association.
15:52Rabid fans. Yes, but very diverse, right? There are some who want the metrics, they want to know
15:58every play and how does that rank versus other players in history. And then there are some that
16:03want the drama and they want to know about any fights or they want to know about injuries. And
16:07and it's, you know, now with the Bundesliga app, you can actually drive up engagement because you
16:13are personalizing the content that's delivered to the user based on what they want. And that's what's
16:19really exciting is it's a personalized fan experience. And they're seeing engagement go
16:22up significantly. The time that the app users are spending there going up because it's the stuff
16:28they are interested in. In terms of just pure connectivity, the fact that the drivers can
16:32speak to their pit crews and speak to each other on the end. And, you know, we at times we can hear
16:37that that definitely adds to the drama. Michelle? No, I was just going to add the tech piece
16:41associated with that. You know, we use something called an unsupervised clustering algorithm to
16:46have the machine basically look at the data in a way that humans don't. So to your facts reference
16:51where you used to sit and do the analysis yourself that, you know, we're now allowing the machine to
16:57train itself to understand who those audiences are and then using Gen.AI and Bedrock and in this case
17:05SageMaker too, also an AWS product to basically name those segments. So something that used to take
17:12marketers, you know, in some cases weeks to do the analysis to name the actual segment, we're now
17:19letting the machine do it for us. And then it's the when it names the segment marketers are then able
17:25to act on it and push that data downstream to do the outreach, whether it be via mobile or social or
17:30app, whatever it might be, the machine is now doing that for you. And it's closing the loop via ML
17:36operations to continue to train it on what those fans want. The loop or the, you know, where the line
17:43ends between the machine learning, the training and the generative AI, that is going to be a science
17:48going forward for any number of applications. And I think this is a really interesting one.
17:54Michelle, are there anything? So if I understand right, Converge has been, it's been in the works
17:58for a while. It's been active in the marketplace for about a year. Anything, you know, any top line kind
18:04of something that surprised you about, you know, the behavior of sports fans or what, you know,
18:10what motivated sponsorship or, you know, things that maybe things that weren't, you know, obviously
18:15intuitive? Well, so first, Converge Consumer was the first Converge. So we took that same data fabric
18:24that we talked about around retail and consumer product data, and we learned from their general
18:29behavior and share of wallet, right? So we took all our learnings from retail, and then we applied
18:35sports specific data around ticket purchase, as well as team affinity based on primary research or
18:44digital research, and then brought the two together to inform the fandom score that I mentioned.
18:51And so the thing that I find most interesting about fandom is that your fandom is not unique,
18:57right? It crosses, it crosses music, it crosses your retail and buying habits, it crosses all the areas
19:05of your life. So you know, think about that, that's intuitive. Yeah, like sports fans are not, you know,
19:11they're not a monolith. They, they are, they are rabid about teams, leagues in different sports. So how do
19:17you look at them holistically as a fan, and then act upon that unique fandom in that audience of one?
19:24Just add to that, Cynthia, because it's, it goes beyond a personalized experience. It's actually a
19:31personalized monetization experience. So you can think about doesn't like that, right? You've got,
19:39you know, subscribers, maybe businesses that have a subscription service, or businesses that are
19:44selling certain content or an ads based business. And you can now with the fandom score, be able to
19:49apply that score or any of those insights provided by generative AI to determine, hey, is this consumer,
19:58do they have, maybe they have a low propensity to subscribe, but they have a high propensity for,
20:04or they have a high ad score. And we'd be able to make, we as a vendor, whomever the vendor is,
20:10could make money off of advertising. So it's okay if they, you know, don't subscribe. Maybe we'll let
20:16them pass through the paywall to get the content because the ad is going to generate more revenue
20:20for us than the subscription, even though there are low propensity to subscribe. So thinking through
20:25how to use that type of information in the monetization model, and to do that in real time
20:30based on the person, I think is really valuable. The streamers for sure, the ARPU on the ad supported
20:36is, is always stronger, even though it has become in the last 10 years, the commercial free has been
20:43seen as more prestige. But if you're talking ARPU, you definitely, you definitely want the ads. Let
20:48me ask you, I'm curious about the score. Is it, is it a numerical score? Is it a certain, is it like
20:53certain traits about a person? How do you, when you say the fandom score, how do you exactly calculate
20:59it and how do you express it in a way that marketers can interpret? It does. It expresses itself
21:03as a number one to a hundred. Um, and it's always in relationship to the brand you're talking about
21:12and the sport or league you're talking about. Right. So, um, but the good news is that you can
21:19adjust it. It's obviously a weighted metric. You can adjust it based on the behaviors that your sponsor
21:25might be interested in. Right. So if a sponsor is interested in selling more live event tickets,
21:30then you can weight it based on discretionary income available to purchase tickets. And obviously
21:37an F1 ticket is a much more pricier spend than a Dodgers game. So you can look at that in a way
21:44that is, you know, that downstream enablement and enactment of, you know, the score as the,
21:52you know, kind of numerical representation. And I would imagine, although with, you know,
21:56some 300 million people in the, in certainly, certainly in the U S monitoring those scores
22:01would be, would be challenging, but I would bet that the, the concept of that, that people are,
22:07would are very eager to look at and see what, what those scores are for folks. And I would imagine
22:12that if you're really high in the fan engagement, you're going to have other attributes. You might be
22:16an early adopter of technology. I would imagine that you're able to draw those kinds of parallels.
22:21Absolutely. The, the algorithm basically surfaces the attributes that are associated with
22:27certain behaviors, right? So this attribute drives live event propensity, right? So it tells you,
22:35and we also use the Gen AI to say, this is a high intending population who care about, to your point,
22:43early technology adoption. So give them an immersive experience that looks different than,
22:48you know, somebody who might be cost conscious, who does not want a technology component, right? So
22:53the, the ability to, to give those, um, I'll say in-person activations in a way that, um, the sponsor
23:01cares about is important to CMOs everywhere. I'll add one more thing that's with the emergence of
23:08agentic AI, where it's in some cases, it's not eyeballs. It's, it's actually machines. You know,
23:14how do you know whether it's a machine or a person and then what data, if it's a machine
23:20scraping to get an insight to, to give to someone versus it's a person that's actually engaging and
23:25wants to spend time on the content. Yeah. Yeah. So how do you, you know, we're, we're using AI to do
23:31that. And actually I was on a panel yesterday with a couple of our partners, um, and Adobe had
23:36mentioned that they've been measuring the number of machines that are, uh, interacting with their,
23:42their, uh, platform. And in the last six months, it went up something like 3,500%. Maybe don't quote
23:49me exactly on the number. It was over 3,000, um, of machines versus people. The machines are doing
23:54something useful, right? They're getting information and sending it somewhere to someone who's going to
23:59use it. But how you present content for that is very different. How you monetize that is very
24:04different. What you have, what is so great is that you're so close to the source of the fan. If it goes
24:10through a machine, through several filters, it's not going to have that same resonance. I think
24:15that, you know, cause something is going to get lost in each translation. So, but that's interesting.
24:19You're using AI to ferret out the machines that are confounding the AI. That is the, that is very
24:25indicative of our times. Yeah. It's the future, right? Agents acting on your behalf is the evolution
24:32of AI, gen AI that are, you know, doing schedule optimization and always looking at the scores,
24:37making recommendations, running in the background. Those algorithms and agents are, are the future
24:43of how fans are going to engage with historically, maybe it was statistics or it was, you know, any
24:49sort of, um, live event even, right? How you're engaging with, um, the device at all times is going
24:56to radically change in the next year. What's cool about agents, especially for in this space,
25:02is it, they can allow you to run multiple campaigns at once, do a B testing way faster
25:07than you could before. Um, I mean, they can parallelize the entire workload. And so, I mean,
25:12I think kind of traditional AB testing, maybe you can do two tests per week, you know, per analyst.
25:18And now with agents, I mean, that number can grow exponentially and then you can, you know,
25:23you can figure out which of these campaigns or which of these experiences is actually driving
25:27better results because the agent was actually able to modify the campaign that you were going to put
25:33out there, test it, get the data, give you the results, and then you keep pushing.
25:38And all that we're talking about, which is again, so personalized, so individualized,
25:43it does underscore the need for massive, massive amounts of cloud storage. It really does because
25:50somebody has to maintain that you, and, and I know, you know, there's costs to maintaining that.
25:55There's a, you have to keep it. I understand you have to, it's something that you have to keep it
26:00fresh. It can't get too stale. Otherwise you, it compromises some of the efficacy of the.
26:05The data challenge is probably, you know, step number one is you have to have a data foundation.
26:10If you look at, and I don't want to say legacy media companies, because it actually plagues every
26:15single industry, frankly, is the data is not all in one place. So if you're thinking about
26:20personalization by aggregating data from the ads team, from the marketing team,
26:24from the content team today, many of them still sit in silos. The functions are in silos.
26:30Their data is in silos and the tech is also siloed. And so being able to bring that together
26:36is, is step number one is getting your data in order, put it all in one place, being able
26:42to use these tools that can give you the insights is really key. And so that's why I actually,
26:47I love the name converge. I'm not a marketing professional or brand person, but I think you guys did a great
26:52job with that name because that is what it's all about. It's bringing that data together to
26:56converge an insight.
26:57And with large organizations, we've all been there in meetings that the, the, the, the possibility of a
27:03small communications, you know, a lapse of communication or somebody didn't talk to
27:08somebody and you're looking for something. And then a week later, somebody is in a meeting
27:12saying, Oh, I have that right here, but you couldn't put your hands on it because it was all in
27:17different, in different spots. That's actually the, that's the number one bottleneck in AB testing
27:21is the analyst doing the work, doesn't have access to the data. And it, you know, the person that
27:26needs to provide them the data hasn't checked their email and all of that. And so the idea of getting
27:30your data in one place and having a comprehensive data strategy, which we do constantly with Deloitte
27:36when we support our customers is, is really step number one.
27:40Yeah. And we've gotten to the point where, you know, prompt engineering, it's a misnomer for sure.
27:45It's just how you interact with the model and ask the questions has made it so that, you know,
27:50when you can't find that piece of data or research, it now spans your organization,
27:55scours it for you, and then surfaces it with an insight or an action you now need to take,
28:00as opposed to having to crunch the number and read the facts.
28:04And something else that kind of blew my mind this week was somebody was saying it very true that,
28:08that learning how to write the best prompts, the most effective, the most,
28:15you know, to really getting across, getting, training your agent and getting your, all of
28:20your AI tools to understand what you want. Like that is the new, just like it, you know,
28:26went from writing term papers to creating PowerPoints, like creating the best prompts and having
28:31the, the strength of the writing of that is going to be like anything that the strength of the quality
28:36of anybody's editorial is, is about how well you can execute it. And, and those prompts at first,
28:42I thought it was at the first, I thought that, that, that, that was like, you know, almost a joke,
28:47but I realized like, that is really going to be significant in the future, really understanding
28:53just as, just as people that understand how code now have a skill that coding of the future is going
28:58to be prompts. There is the prompt side of it, which is a reactive engagement mode. So like I,
29:03I put in a prompt and you know, the, the LLM is reacting and giving me an answer, but the,
29:10the agentic kind of wave that is, is upon us is just proactive. It's just doing things on our behalf.
29:17Circusing it. Yeah.
29:18And that to an end goal.
29:19I'm still wrapping my mind around, you know, there will come a time,
29:23everything is becoming Hollywood and Hollywood. Everybody has an agent soon. Everybody will have
29:28an agent and somebody will have to store all that information. Or maybe, maybe the state of that
29:33art will be finding more efficient ways to store it in a way that you can put it over here, but grab
29:38it when the person, either the person needs it or the person putting in the prompt.
29:41The cool thing about agents is they're kind of optimized for a task or a part of a workflow.
29:47And so if you think about someone in a marketing function, you know, they've got to think about
29:51multiple workflows and processes, and you can have agents that support a subsegment of that.
29:56And so having them all work together on behalf of that individual to figure out, Hey, what is the
30:03most optimal, if you're optimizing for cost, or if you're optimizing for reach or whatever the
30:07outcome you're trying to achieve, multiple agents can come together. We have multiple partners in our,
30:14our AWS partner network, and they're all contributing agents to a broader network because it no,
30:21you can't do it with one company. I think it does require partnerships and
30:26experts in certain verticals or in certain processes or in certain workflows and bringing
30:30all those agents together to interoperate and deliver that optimized process. Um, I think that
30:36is going to be the future and it is all proactive. Like they're doing it on their own and the business
30:42leader's job is now to ask the right question and what the key outcome is that they're trying to achieve.
30:47Knowing who to go to, who's going to bring me that superpower. That's so, that's so interesting.
30:51That is going to change the needs of business. That is really something, Michelle. Again, I know
30:56it's still early days yet, forgive me early innings, but do you have any examples of, of cases where
31:02you've been able to, you know, the information, the data analyzed has been able to, uh, give a sponsor,
31:08you know, a really, really successful placement. Like you just hit the sweet spot of where they,
31:13where that brand or service wanted to be. So I think everybody is gearing up for a world cup
31:19coming to the U S next year. Right. And I think more and more folks are interested in
31:25a North American market of who are the people, what do they care about? What are their buying
31:31habits? I recently had an international company reach out to me and asked me, what do people in
31:37Kansas city really care about? I was like barbecue maybe I don't know. I'm kidding. Um, but it's a
31:45great example of using targeted marketing for the purposes of all eyes are going to be on 12 North
31:52American cities or 16, I think North American cities in a way that to be honest has never been
31:59the focus of marketers. And now what fandom enables is what my fandom looks like in LA is different than
32:06someone's fandom in Kansas city. So in Kansas city, they stay till the end of the game.
32:10Yeah, exactly. Everybody in LA is trying to beat the traffic, but, um, but that's like an interesting
32:16marketing example of how do I engage in these 16 different markets in a way that, you know,
32:23looks different than it would as a global brand in a mass market sense. I think ultimately next year
32:29will be an interesting kind of test case for all things, personalized outreach and engagement.
32:35And to your point, I mean, you know, marketers, marketers that are based overseas, you know,
32:41even 10 years ago, wouldn't have had a 10th of the tools that they can get now geo-targeted to
32:47Kansas city. My, my first thought is Kansas city, Kansas or Kansas city, Missouri.
32:52By the way, I love that Michelle's like, um, the world cup is a test case.
32:56It's like, let me ask you, Michelle, is there any, um, next frontier for conversion or you've done
33:05so that the, the, the system has been applied to retail and sports. Is there, where are you going
33:10next? So the next frontier is media, of course. So converge media, not too far off, but applying the
33:16same concepts, right. Of measurement engagement. What does it mean for media organizations who care about
33:23all things fandom for potential subscribers and, um, who is next as a subscriber and who doesn't
33:30subscribe today? And how do I talk to them in a, in a tone of voice that gets them down the funnel to
33:36convert faster? I think your fandom in characters, titles, worlds, all influences your purchasing behavior.
33:45Um, when you think about streaming and subscription. And so for us, that's the next frontier to take all
33:51that we've learned and done with AWS is to apply to yet another industry for us.
33:58And I wanted to ask you, is this largely us centric at this point? Are you largely focusing on the us
34:03consumer today? Our data fabric is focused on the us consumer. We definitely have, um, ambitions on
34:08global mark, more global markets. Um, but as you know, data around individuals is variable by country.
34:16So our ability to stitch it together in a fabric, um, is going to take a little work, but yeah,
34:21we're definitely working on global and other industries as well.
34:24Aruba, anything, um, anything else you want to say about converge or anything
34:28that's really important among the themes that we've been talking about, identifying fandom, finding that
34:34moment of opportunity, anything on the horizon for AWS that you'd like to talk about?
34:39Yeah, I think it's all about, you know, bringing together the different capabilities and then
34:43providing new experiences. And actually, you know, Cynthia, I was speaking with you and Michelle
34:47earlier about something I'm really excited about that we're showing casing here in the Amazon port,
34:52and that's taking multiple different large language models. So Amazon Nova's family of large language
34:59models that are optimized for, you know, speech to text or text to analytics or creating an image or
35:05creating a video. Um, and you can go in and have an experience where you tell this machine what inspires
35:13you, what are the fragrances that you like, or what are notes that you like? And then it'll create an ad
35:20campaign for your personalized perfume with a name, with an image, with a video. And then actually there's
35:27a perfumer there that will make it based on the notes there. And so think about how generative AI
35:31is providing personalization across all aspects of our lives and taking something that isn't just
35:37this, it's all in the cloud. We don't know what that is. It's, you know, big data or data analytics,
35:42machine learning agents, it makes it something tangible that you can touch. And that's what I love
35:47about the possibilities here is it's giving you things that you can only imagine. And actually you
35:51didn't imagine it, the machine imagined it and made it real for you. I would, you know,
35:56the idea of creating your own perfume and to the point of even having an ad like that,
36:01you know, that gives it a whimsical slogan and a name it's, but I mean, it just takes it all the
36:06way through. And you can think about like what AI is providing is democratization of all of these
36:12capabilities that used to require, you know, deep expertise in all these different areas. And now
36:17here it is, and it's deeply personalized to the user.
36:20The ability to train it to exactly your agents or your large language models to exactly what you
36:27need. I think that that is, we're just, just starting to scratch the surface of that and
36:32certainly in media and entertainment, but just in personal living our lives. So I really appreciate
36:37you both spending the time here. I'm going to listen to this twice because there's a lot,
36:41there's a lot to learn here, but I really, I mean, you couldn't be more in the trenches of it.
36:46Also, I think conversations like this express that there is so much, there is so much opportunity to
36:52unleash with new technology. It is not something to fear. It's something to harness in a way that
36:59you've given us some great use cases. And listeners, since you can't see it, I just want to say that
37:04we have been talking here a lot about clouds. There is not a cloud in the blue sky here in Cannes,
37:10but only in Cannes would we have this deeply technical conversation on a yacht where the
37:16yacht next door, they're setting up for a party with disco balls and discombobulated heads.
37:22That's Cannes for you. Thank you all so much. Really appreciate it. Really appreciate your time.
37:27Really appreciate your thoughts. Thanks for having us. Thank you.
Recommended
1:16
|
Up next
2:30
2:22
46:06