- 6 months ago
From Search to Chat: What Happens When Your Customers Talk to AI?
Category
🤖
TechTranscript
00:00All right, what a wonderful panel we have here.
00:03I look forward to learning a lot from these folks.
00:06Good morning, how are you?
00:08Good morning, good. How are you? How's everyone doing?
00:10Good? Good.
00:12I've had like four croissants and four cappuccinos, so I'm good.
00:16All right, I've taken three modafinils.
00:19Let's get going.
00:21First, I just want to, before we get into the real meat,
00:23which is explaining how marketing changes when nothing's real,
00:26as the introduction just went,
00:28I want each of you, I'm going to go rapid fire,
00:31to just tell me a data point or a number
00:34that indicates how much marketing is changing because of AI.
00:39Something that you've seen in the statistics
00:41that shows what's exactly happening.
00:43Esmita, you're smiling the most.
00:45Maybe it's just because you work at L'Oreal and you look so great,
00:47but tell me what that number is.
00:49No, no, I'm smiling because...
00:51By the way, isn't this like a strikingly handsome panel here?
00:53Anyway, go ahead.
00:55You know, I'm smiling because I'm going to say a strange number.
01:00You know the name of the panel is Search to Chat,
01:03and the number I'm going to give is zero,
01:05because once you think that there is a zero-click search,
01:10you know that AI is here and here to stay.
01:13So first number is zero, and that's on a lighter note.
01:16But for us, it is important consumer engagement with beauty,
01:19and what we see is in that consumer engagement,
01:23when you drive awareness, AI-powered media gives efficacy.
01:26When we are into consideration, we can drive average baskets.
01:30When you are into purchase, the conversion goes up,
01:33and at the end, when we use augmented care agents,
01:38consumer satisfaction goes up.
01:40So the number goes up at each part of the funnel.
01:44Caitlin?
01:44Yeah, the one that sticks with me the most is one,
01:47just 81% of consumers want to interact
01:50with a highly personalized digital...
01:53Wait, 81%?
01:5481.
01:5581%.
01:56But when that is delivered successfully,
01:59we see for our customers the lifetime value increase
02:03from anywhere from 10% to 30% with that brand
02:06when they deliver it with trust and personalization,
02:09and that just shows that AI and marketing is working.
02:1181% want to interact with AI, not with a human?
02:15Yes.
02:16So 19% prefer the human?
02:17They want highly digital, personalized, frictionless experiences.
02:22So what that means to a consumer,
02:23you shouldn't have to care if,
02:25hey, is this an agent?
02:27What is this, a bot?
02:28You just want the brand to know you,
02:30to understand your preferences,
02:32your interactions with the company,
02:33and if you can do that successfully,
02:36we see customer lifetime value increase from 10% to 30%.
02:39But in general, from what you're seeing,
02:42do people prefer interacting with people or with AI agents?
02:46Well, we think the secret sauce for success,
02:48depending on the use case, Nick,
02:49is this blend of humans and agents together.
02:52Yep.
02:52So I think we can all shake our heads and nod
02:54when you interact with a brand.
02:55Anything that's just frictionless,
02:57whether it's a human or an agent,
02:58you want them to know you,
02:59and you want them to know your preferences.
03:01But normally, it's just speed.
03:03And so 81% are actually preferring
03:06that high level of personalization,
03:08and I would say digital self-service experiences.
03:11But if it does require the handoff into a human engagement,
03:17they want it to be persistent,
03:18and you to still know that history
03:21that you might have been having
03:22with the digital self-service experience as well.
03:24Wow, very interesting.
03:25Dave?
03:25So AI has always been very important at Meta
03:28and actually relevant to Paris.
03:30We opened our FAIR AI Research Lab here in Paris in 2013.
03:34So it's been a part of our business for a long time.
03:37My stat to your question is 50%.
03:39So if you think about your Instagram feed,
03:42over half of that content in your Instagram feed
03:45is now recommended by AI.
03:47And that's an incredibly complex math problem.
03:50So you think about billions of people using our services,
03:52so 3.4 billion people a day use one of our services,
03:57times tens or hundreds of millions of pieces of content
04:00that brands like L'Oreal create.
04:02The math problem of matching that
04:04and determining what's relevant for one person
04:06is incredibly complex.
04:08AI powers that.
04:09And then the question is, are those recommendations good?
04:12And I think one of the real tests is,
04:14if the recommendation is right,
04:16it might be so good you want to share it to a friend.
04:18Has anybody here ever shared a reel to somebody on Instagram?
04:21You sent it to a friend, yes?
04:22I do this too.
04:24Four and a half billion times a day,
04:27someone sees a reel and shares it to a friend through messaging.
04:30So that linkage of our feed services and messaging,
04:34that's where marketers can really take advantage.
04:37Wait, you said 50% comes from AI?
04:39I would have said 100%.
04:41So there's AI involved in everything that gets selected
04:45because there's so many permutations that you could come up with.
04:47The change for our services was,
04:49if you used Facebook or Instagram five years ago,
04:52you would have had your friend graph
04:54and you would have seen only content from friends
04:57or from accounts that you follow.
04:58And that alone was quite complex.
05:00What we started to do is realize that
05:02there's so much other content from brands,
05:04from creators that people want to see.
05:07and if you introduce that content to them,
05:09that increases satisfaction, time spent,
05:11all the metrics go up.
05:13But since you've opened that aperture
05:15of how much content is out there,
05:17the AI complexity of that challenge is much bigger
05:19and you can only do that with better AI.
05:21All right, Azmita, let's talk about zero-click searches.
05:24Let's talk about a new thing you've launched.
05:25You have a system.
05:26It's called Beauty Genius.
05:28You go into it.
05:29You get recommendations.
05:30It told me to brush my hair more frequently.
05:32Tell me about it and how people should use it.
05:35Yeah.
05:36So L'Oreal Paris is the number one beauty brand in the world.
05:40And L'Oreal Paris has just launched Beauty Genius.
05:44And Beauty Genius is an agentic AI-powered personal beauty assistant
05:51that can remain 24-7 in your pockets.
05:54What Beauty Genius does is gives you personalized diagnostic,
05:59personalized recommendations, personalized Q&A,
06:03and then you can actually click and buy the same thing
06:08because it links to Amazon or Walmart or Target.
06:11And what's very, very interesting is our path-breaking partnership
06:17with the number one messaging app in the world, which is WhatsApp,
06:22because Beauty Genius will now be available on WhatsApp.
06:26Thank you, Dave, for that.
06:27Thank you.
06:29Well, at the Atlantic, we use Signal,
06:31but I'm happy to have you talk about WhatsApp.
06:34Does Beauty Genius ever recommend products not from L'Oreal?
06:38Like, would it ever recommend one of your competitors' products?
06:42So that's great.
06:44Beauty Genius is today on the L'Oreal Paris USA website,
06:49and it has had 400,000 conversations.
06:52It uses 750 products because the product information, etc.,
06:57has to be right from L'Oreal Paris brand,
07:00skincare, haircare, makeup.
07:02What you will also see on the Vivatech booth of L'Oreal
07:06is an AI matchmaker called Nolly.com.
07:11Nolly.com is a multi-brand marketplace platform using AI,
07:16which then uses different brands and different products
07:20and, you know, can start recommending from there.
07:23And let's just stay on this for a minute
07:25because I think it gets at one of the most important things
07:27that's going to happen in marketing.
07:29So as they said in the introduction,
07:30Google search is going away.
07:32Google will just answer questions.
07:34So in the old days, you would use search engine optimization,
07:37you would buy ads on Google,
07:39and you would drive people to the L'Oreal website.
07:41Now you have to get people into your Beauty Genius.
07:46How are you going to do that?
07:47You can do it in Facebook.
07:48Facebook will drive you there.
07:48Instagram will drive you there.
07:49But how are you going to get them from Google,
07:54ChatGPT, Anthropic, Perplexity?
07:56How are you going to get them into your system?
07:59Okay.
08:00Nice question.
08:01And I just told you that we are starting a partnership with WhatsApp.
08:05Yeah?
08:06That is one way.
08:07Because, for example, you're in the social media feed
08:10and you're on Instagram or Facebook
08:12and you're just looking at things that you like.
08:15If you like something on beauty
08:17and there is the L'Oreal Paris messaging there,
08:21you could click on that
08:23and automatically reach the WhatsApp
08:26and start chatting with Beauty Genius on the WhatsApp
08:30as if Beauty Genius was your friend.
08:32When you have WhatsApp,
08:33you see the photographs of your friend there.
08:36So you see the photograph of Beauty Genius there.
08:38So the journey of I want a deeper one-to-one conversation
08:42can go from there directly into the conversational AI,
08:47which is where we see a lot of conversations coming
08:49and therefore more time spent
08:51and therefore a higher conversion on Amazons and Walmart of the world.
08:55Okay.
08:56Kaylin, let's talk about some of your systems,
08:58your agent force.
08:59What are some of the surprising things that you have learned?
09:02You've put these agents out on the world.
09:04You've got AI bots everywhere.
09:06Mark Benioff has said he's more or less done hiring humans.
09:09What are you learning from all those agents you've got out there?
09:11There have been some fascinating insights.
09:14So we just launched Agent Force.
09:16It's just been about two quarters ago.
09:18Now over 8,000 customers.
09:20And one of the things that I've found most fascinating
09:22is we wanted to be customer zero.
09:25We wanted to drive our agentic transformation from within
09:29to learn patterns that are working and not working
09:32along with our customers.
09:33So we put Agent Force, an AI agent,
09:36right on our website from the beginning.
09:38We have 72 million hits on our website every single year.
09:41with questions around products or acquisitions you can imagine.
09:45One of the most fascinating insights was if you look at
09:48when we first put it on our website,
09:50people would ask very simple questions, right?
09:53How do I turn this on?
09:54How do I change a password?
09:56Once we developed trust with them,
09:58we knew who they were,
10:00what their preferences were in a trusted fashion.
10:02we saw the behavior change.
10:04Meaning the level of questions they would ask
10:08once they had that frictionless, trusted,
10:11personalized experience,
10:13they started going deeper and deeper.
10:14So we started tracking the conversations
10:17that were happening with our brand month one,
10:19month two, month three.
10:20And what we've seen is that trust
10:22allows you to go further, faster with agentics.
10:26but what's powerful is what's the business case
10:28that this is unlocking for our customers.
10:30That gets back to the opening stat
10:32of really 10 to 30% increase
10:35in their lifetime value with us
10:37because they want that level of personalized support.
10:40So it's been interesting to see
10:42the behavior shift with time
10:45and that trust drives curiosity.
10:47And for some customers,
10:49this is another one,
10:49new workloads, new workloads.
10:52So you'll start to ask things of that brand
10:54maybe you wouldn't have asked before.
10:56So we've got a large retailer in the United States
10:58say, hey, what, how many gift cards
11:01do I still have outstanding?
11:02What's the balance?
11:03Those were things that their brand
11:04wasn't being asked before,
11:06but now was starting to become asked.
11:09Once first, it was a frictionless, trusted experience.
11:11So you go further, faster.
11:12So if you're a company,
11:14you should recognize
11:15that once you have agents answering questions,
11:17people will ask more and more detailed questions.
11:20So you can also nudge them
11:21to more and more useful patterns
11:24for them and for you.
11:25Absolutely.
11:26Okay.
11:26But to get there,
11:28it's 100% critical that you have trust
11:31because we also have all interacted with a brand
11:34where you don't have that good experience,
11:36you never go back.
11:37Dave, did you want to jump in there on trust,
11:39either for or against?
11:40So when we think about messaging experiences,
11:44we have a Lama model that we've built
11:46that we use for different purposes.
11:49We offer it to the market through open source
11:50so a company can take it and use it for their own purposes.
11:53We also use it for our meta AI tool,
11:56which is embedded inside each of our apps.
11:58So we have roughly a billion people per month
12:01that use that search feature in WhatsApp
12:03and Instagram and Facebook.
12:06once you have that conversation
12:09and you can get someone into a messaging relationship,
12:12a one-to-one message,
12:13there's so many things that you can do.
12:15So as Mita talked about Beauty Genius,
12:16but you can think about customer service,
12:19getting information, upselling, cross-selling.
12:22The trust piece is critical.
12:24But I think the main consumer behavior
12:26that we see in messaging is
12:29people want to communicate with the business
12:30just like they would communicate with a friend or a family.
12:33If I have a question, I'm going to text somebody.
12:35That's what I'm going to do.
12:36I'm going to do it on WhatsApp.
12:37We see 600 million conversations per day
12:41between one of our users and a business.
12:43So people want businesses to communicate with them
12:47in a natural way.
12:48If anybody here is a teenager at home
12:49and you send them an email,
12:51do they read your email?
12:53No, because they don't use email.
12:55They want messaging.
12:55So it's even more prominent for our younger audience base.
12:58But you have to deliver on those expectations.
13:01It can't be slow.
13:02There can't be latency.
13:03People really expect that the speed
13:05of those messaging experiences will be excellent.
13:08May I add something to that?
13:10I don't know if the audience has it in mind
13:12because we are talking about meta, WhatsApp,
13:15one-to-one messaging.
13:17Then we are talking about Salesforce
13:19and Salesforce has this history
13:21of making one-to-one relationships with CRM
13:24and all, you know, for all of us.
13:27Here we are experimenting with both WhatsApp and Salesforce
13:31together with L'Oreal brands in our market in Brazil.
13:36So in a market like Brazil, it's a very social market.
13:39And there is a lot of trust in these social interactions one-to-one.
13:44So what happens in Brazil as a market is
13:46we start having conversations on WhatsApp,
13:50which can then be either inbound, they are coming to us,
13:55or outbound with a CRM,
13:57you know, Marketing Service Cloud of Salesforce.
13:59So it can be inbound or outbound, but you're chatting on WhatsApp,
14:04and the consumer is able to have a deeper conversation
14:07thanks to all these technologies coming together.
14:10And that is a big portion of our direct-to-commerce website
14:14in Brazil as a market.
14:17And we are really scaling this social commerce.
14:19And I'll give Asmita and her team a compliment that
14:22you think about markets like Brazil, Mexico, India, Indonesia,
14:26where there's incredibly high usage of WhatsApp
14:28and sort of how these countries run.
14:31It's great to have those early experiments
14:33and test something and see how it works in one market.
14:37Asmita has the challenge to reach billions of consumers
14:39and expand her business.
14:41That's your expectation from your CFO, your CEO, I'm sure.
14:43Yeah.
14:44So it's how do you get those early experiments
14:47but then find a way to scale them
14:48so they really impact the overall business?
14:51And I think that's one thing that L'Oreal is...
14:52Let me ask, okay, so the world used to work
14:55where it was humans and they were talking to humans, right?
14:58And they would call your reps and they would talk to people.
15:01And now we're in this very brief moment
15:03where it's humans and they're talking to AI agents, right?
15:06And they're talking to your agents and your agents
15:07and your agents.
15:09But soon it's going to be agents talking to agents, right?
15:12You're not going to have humans who are coming to Salesforce.
15:14I'm just going to send out my agent
15:16and I'm going to go to my agent,
15:18find me some skincare stuff, right?
15:20And I'm not going to go look at it.
15:22I'm going to go mow the lawn,
15:23play basketball and learn guitar, right?
15:25So what happens when you have agents talking to agents?
15:29Are they even going to speak English in AgentForce
15:31or is it going to be just a new language?
15:33Well, it's interesting.
15:34I don't know if anybody saw the recent Gartner report
15:36that a lot of teams, it was 50% of companies
15:39have said they're going all in with digital workforces
15:42are now actually kind of dialing that back
15:43because there is the secret sauce,
15:46the blend of humans, physical labor and digital labor
15:49working together for trust
15:50and maintaining that persistency
15:52wherever you're interacting with the brand.
15:54But to your point, when we get to this place
15:56where you've got your personal agents
15:58for Kaylin or Nick
15:59now interacting with L'Oreal,
16:01buying me the best skincare products
16:03I can possibly find,
16:05what does that unlock?
16:06and what I get excited about for that
16:08is it unlocks the greatest kind of scarce resource
16:12for all of us, time.
16:14And it unlocks knowledge.
16:16And what do I mean by that?
16:18We are gated by time.
16:19If you are trying to find the best skincare products,
16:22you might have five minutes.
16:25Your agent can do that at scale and depth
16:29and give you the time back for all of us,
16:32which we could all use more of,
16:34but also not just maybe the top three links,
16:37which is the history of search,
16:39but now truly unlocking, democratizing knowledge.
16:42So it really is getting you the best that's for you, Nick.
16:45And so what's exciting
16:46when I think around that agent to agent
16:48on both the consumer and business side of the house
16:50is it is neutralizing both time
16:53as well as intelligence
16:54for anyone who's leaning into that level of innovation,
16:57which is exciting.
16:58But okay, I like it.
17:00And that's saving me time.
17:01And that's why I'm off playing guitar,
17:02which I don't do enough now.
17:03I'm doing 20 mile runs
17:04like he's out doing this morning in Paris.
17:06Well, not while I'm playing guitar, but yes.
17:08So more time for both of those would be amazing.
17:10But as a marketer,
17:12how do you make it so that my agent
17:14finds L'Oreal's products
17:16and not somebody else's products?
17:17Like you have a whole mechanism for advertising
17:21when it's agent to agent.
17:22What's like, how is that even going to work?
17:24Like, how do you buy an ad
17:26to increase your brand halo to an agent?
17:32Well, that's where Salesforce
17:34has always been a customer company.
17:36So wherever you want to interact
17:38with somebody's brand,
17:39and you can both speak to this as well,
17:41that they understand
17:42whether you're interacting with them digitally
17:44or physically or across what L'Oreal is,
17:47what, 46 global different brands,
17:49far more, but just global brands,
17:51that they're capturing that profile
17:53so that any opportunity to interact with you
17:56isn't missed, right?
17:57That we're unifying that view of the customer,
18:00no matter brand, what market,
18:01and then advertising to them accordingly.
18:03But I'd love to hear your thoughts.
18:06I can add to what Kaylin is saying,
18:08and you tell us what you think, Dave.
18:11And I'm going to quote here
18:13another one of our platforms like Google,
18:15you know, which is saying that consumers today
18:17are scrolling, streaming, shopping,
18:23searching, all at the same time,
18:25anywhere, everywhere.
18:26And this job can be done by agents
18:28is the question, you know,
18:30that we are asking to ourselves.
18:33So, sorry, I lost my thought
18:35because what I was going to say on this
18:38about agent-to-agent is
18:39we just spoke about the whole topic
18:42of this panel, search-to-chat.
18:45When you're talking search-to-chat,
18:47there is this thing about SEO,
18:49search engine optimization,
18:51moving to generative engine optimization,
18:55SEO-to-geo.
18:57What does SEO-to-geo mean?
18:59It means still all the information
19:02that is there on our brands
19:04in the World Wide Web
19:05that has to be structured in a way
19:08that it can be picked up by AI.
19:11Today, it is structured in a certain way
19:14on the websites.
19:15But the meta-tagging,
19:16are the photographs tagged,
19:18is the information consistent
19:20so that when the AI search
19:23starts to pick up information,
19:25it is able to read that information well
19:27and therefore pick it up.
19:29So the task at our end
19:31becomes that our brands
19:33and their presence
19:34and their footprint online
19:36is more geo-friendly.
19:39And we are starting that work.
19:41It's a hard work.
19:42So, okay.
19:43So will your ads work like that?
19:45So L'Oreal buys these beautiful ads
19:47on the Atlantic.
19:47They look lovely.
19:49That's for humans.
19:50We're going to start setting
19:52our Atlantic site
19:53to have a version for humans
19:54and a version for bots.
19:55Will there be L'Oreal ads
19:57embedded in the metadata
19:58of the version of the Atlantic
20:00that is read by bots?
20:03Yes.
20:04Amazing.
20:05Possible.
20:06Yes.
20:06We'll close the deal
20:07in the back room.
20:08Thank you.
20:08Yeah.
20:09So, Nick,
20:10I would just add that
20:11once someone's involved
20:12in a message thread
20:13or they're engaged
20:14with an agent,
20:15that's great.
20:16But you have to get people there.
20:17They're not going to find it
20:18on their own.
20:19So we think a lot about
20:20Discovery Engine
20:21of at the top of the scale,
20:23how are you producing
20:24the right content
20:25that reaches people,
20:27that triggers that stimulus
20:28for them to take that next step,
20:29to go to Beauty Genius,
20:31to get involved,
20:31to reach out?
20:33And what we've seen at Meta
20:35is you really have to have
20:36a broad enough range of content
20:38so you're finding the right people.
20:39So creators is one tactic
20:42that's been very effective
20:43and it's the old model
20:45was you'd have one campaign
20:46that ran for six months.
20:47That was the campaign.
20:48It was a 30-second spot.
20:49That was it.
20:49And the new model now
20:50is you need a range of content,
20:52particularly from creators
20:53that meet different types of needs
20:56and then our algorithm
20:57can find that content.
20:59So we think about this concept
21:00of creative is the new targeting,
21:02but you have to have
21:03enough range of content
21:05to deliver.
21:06So advertisers like L'Oreal
21:08and their use of creators
21:09not necessarily just
21:10the biggest creators
21:11that have a huge following,
21:12it might be niche creators
21:13that have a smaller,
21:14more particular interest
21:15and very effective.
21:16And if creative is the new targeting
21:18and I'm just throwing it out there,
21:21then that creative
21:23and that targeting
21:25can be more intelligent
21:26around when it is talking to me
21:29and when it is talking to my agent,
21:31it can be different creative.
21:33Right.
21:34Because the way they process it
21:35is different.
21:37So maybe we'll go there.
21:40And it's the learning loops
21:43that are adjusting in real time
21:45that are powerful to see
21:46so that you're turning all of this,
21:49what can be kind of digital exhaust
21:50that's been lost into fuel.
21:53And when you can capture that in real time
21:55and I'll give you some,
21:56you like that?
21:57I like that.
21:57I like that.
21:59I can give you some examples.
22:01Right.
22:02When we work with customers,
22:03maybe there is some new trend
22:06or theme that is coming up consistently
22:09and that's within their sales organization.
22:12But if their agents are saying,
22:13let's say they have a sales agent,
22:14but then their marketing agents can see,
22:18this is a change in real time.
22:20We should start making a campaign effectively
22:22because we're having a lot of demand come in
22:24from this side of the house,
22:26this cross-functional organizational intelligence
22:28that when you start having an agentic platform,
22:32all again, untrusted data,
22:34it can drive omniscience
22:37so that as one part
22:38or one signal of the business
22:39is learning something new,
22:42all of the other agents
22:44then are learning as well,
22:45which is powerful to see.
22:46Can I ask whether there might be
22:49a little backlash
22:50and a marketing opportunity?
22:51So one of the interesting things
22:52we've seen at The Atlantic
22:53is that after the election,
22:56a higher percentage of people
22:57started to ask for print magazines
22:58and we were like,
22:59that's strange.
23:01And so we surveyed them.
23:02We said,
23:02why is the demand for print going up?
23:05And they're like,
23:06we don't trust anything on the internet anymore.
23:07We just want to opt out of this nonsense.
23:11And so here on this panel,
23:12we've been,
23:13and I'm fully into the world
23:14where it's agents and agents
23:16and we're selling ads in the metadata.
23:18But for some of the world,
23:20the notion of talking to an agent
23:22or sending an agent to talk to an agent
23:23is going to be unappealing.
23:25So how are marketers going to reach them?
23:27Dave?
23:28You think about a concept
23:29called omni-channel,
23:30which is you might start
23:31a conversation online.
23:32You figure out what product
23:33that you're interested in.
23:35A lot of people will want to go
23:36into a physical store
23:37or will want to talk to a live person
23:39on the phone for advice.
23:40Yeah.
23:41And when that wish is there,
23:43you have to make sure
23:44that your systems are right
23:45so that that employee in the store,
23:48they have the history and the context
23:50that they can see like,
23:51oh, I see the information
23:52you've given us.
23:53Now I can pick that up
23:54and close that sale.
23:55So we think about it
23:56as omni-channel
23:57and making sure
23:57that those in-person staff
23:59are ready to catch that
24:00and treat the customer well.
24:02Yeah.
24:03I think it's a very profound question
24:06and it goes with technology
24:08and culture and people
24:10have a very complex
24:12and a bi-directional relationship.
24:15You know?
24:16On one hand,
24:17technology shapes culture
24:18and our behaviors.
24:19I'll give you an example
24:21and I'll come to
24:22what could be the problem.
24:24Squid games.
24:25It became popular,
24:27Korean culture
24:28and the details of it
24:29became popular
24:30because Netflix
24:31has the technology
24:32to bring it to everybody.
24:35And then culture
24:36can shape technology
24:37because in some cultures,
24:40live streaming
24:41is very popular in China.
24:42It's not so popular
24:43in all parts of the world.
24:45I'm giving examples
24:46that you already know.
24:47So this relationship
24:48is bi-directional.
24:50Given that,
24:52yes, AI is coming.
24:54It's around us.
24:55It's everywhere.
24:56Meta AI has 1 billion users
24:58this year.
24:59So that's huge.
25:01I think it's about
25:02how we use it.
25:04We can make a bold statement
25:06saying that
25:07on one hand,
25:09AI can tell
25:10very meaningful stories
25:13because very relevant,
25:15very one-to-one,
25:16very this.
25:17But on the other hand,
25:18it is true
25:19that we can drown
25:20in the AI slop.
25:22Yeah.
25:22We can all drown in it.
25:24You know?
25:24So how we are going to use it,
25:26what we are going to do with it,
25:29it's in our hands
25:30a bit today.
25:32And I think we have to do it responsibly.
25:36Kaylin?
25:37You know, it's interesting.
25:38We can all probably relate
25:40to this as consumers.
25:42I actually, to your point,
25:43feel drowning
25:44in some of the
25:45proactive marketing.
25:47However,
25:47I don't think anyone
25:49would disagree
25:49that if you go
25:50to try to book
25:51your travel
25:53or if you want
25:54to go have
25:55a personalized experience
25:56that you are initiating
25:58with a brand,
25:59that you want that
26:00to be frictionless,
26:02personalized,
26:03multimodal,
26:03persistent.
26:04And so,
26:05to your point of
26:06people are drowning
26:07in some of the AI slop.
26:09That's true.
26:09Yes.
26:10Yes.
26:11However,
26:12when I interact
26:12with a brand,
26:13I want them to have
26:14all of that knowledge
26:15with me
26:16and then that's
26:16where we go back
26:17to my opening point
26:18where you go further,
26:19faster with trust
26:19and improve CTV
26:21or lifetime value.
26:23So I think it's a matter
26:24of how you build trust
26:25with the brands
26:26and the interaction,
26:29whether it's push or pull.
26:31But I don't think
26:32there's one human
26:33that would say
26:34they don't want
26:35a frictionless,
26:36personalized,
26:37digital experience
26:38if they could have it
26:38with trust.
26:39Right.
26:40Great.
26:40So your frictionless,
26:41personalized experience
26:42will be your lifeboat
26:43to get you out
26:44of the AI slop.
26:45Perfect.
26:46Thank you very much.
26:47Yes.
26:47There you go.
26:47Exactly on time
26:48to the second.
26:48Thank you, Dave.
26:49Thank you, Asmita.
26:49Thank you, Kayla.
26:50Thank you, everybody.
26:51All right.
26:51Well, that was nice.
26:53Thank you.
26:53Thank you.
26:54Thank you.
Comments