- il y a 8 heures
Capturing Attention in the Age of AI
Catégorie
🤖
TechnologieTranscription
00:00Bonjour, je suis ici aujourd'hui pour partager avec vous et parler de l'attention, ou de la fameuse subjecte
00:12de l'attention.
00:13Et plus que ça, notre AI est changement, ou va même changer plus, la fameuse subjecte de l'attention.
00:22So, the first question we ask ourselves is, do you think that AI will be attention's best friend, or frenemy?
00:34And that's not an obvious question, because probably like me, you were thinking, or you are thinking, that obviously, AI
00:44will be the best friend of the attention.
00:47Yes, thanks to AI, we can target better, target further, but also personalise, at scale, all our communication.
00:57That's true.
00:59So yes, it sounds that, obviously, AI will be the best friend of the attention.
01:05But let's look at it a bit further, and you're going to understand that the question is a real question,
01:11and you're going to tell me at the end, what do you think?
01:16First of all, when we talk about the attention, and it is a key point, we are talking about media
01:23impact.
01:24So obviously, the choice of media, which media you choose, and how you use the media is key.
01:32But what we know as Kantar, when we look at all our data, is the creative quality is also really,
01:42really key.
01:43So your content comes for half of the efficiency of your media plan, and attention, and efficacy.
01:54So let's start with the content.
01:56I know you heard about it just before.
01:59I've seen Agathe Bousquet talking about content, and AI, and content.
02:04So let's, I'm not going to say to you how you can use AI to produce content.
02:11I'm sure you know already, and you are all doing it.
02:13But what is interesting is, is it efficient to do so?
02:19The first fear that we can have, or we can see already, is the famous sea of seamness.
02:28It's true that if you use the same AI tool with the same prompt, you're going to have the same
02:36campaign at the end.
02:37Let me take one example that everyone knows very well, the famous campaign for the car dealers' open doors.
02:45You know, the famous one that we all see.
02:48Imagine you are using an AI tool, let's say Midjourney.
02:53You prompt Midjourney with your classical message on the car dealers' open doors.
03:00At the end, you're going to have all the same campaign.
03:02The only difference will be the brand, and probably, fully, the car model.
03:08So the real risk is clearly this one.
03:12That's the first risk.
03:14But there is more.
03:16We all agree that using AI to produce content make us win time and win money.
03:25Yes, it takes less time, and yes, it costs less.
03:29But is it really always efficient?
03:33It's the question, I don't know if you were at the conference before where Agathe Bousquet was there.
03:39She was talking about it.
03:41And it is true that we need to be careful on how we use it.
03:45And what we can share with you at Cantar, because we are doing all these kind of studies every day,
03:52we can share with you that, no, it's not always efficient.
03:56Let me take two examples.
03:58The first one is an example that probably you all have seen.
04:02It's the last Christmas campaign from Coca-Cola.
04:06This campaign has been done with AI.
04:08Let me share it with you, and we can discuss about it after.
04:15Normally there is some sand.
04:19You're going to see it without the sand.
04:32You're going to see it without the sand.
04:36Always coming, always coming, always coming, always coming.
04:42So that's the last Christmas ad from Coke, and it was clearly produced fully by AI.
04:51But it was created by human.
04:56Human has thought that, yes, let's make a Christmas ad which celebrates the heritage of Christmas by Coca-Cola.
05:05And, yes, a human was really driving all the AI tools to make it magical.
05:12You don't make it magical without human.
05:15And when we have tested this campaign, for you to know, this campaign is the fourth best campaign we have
05:23tested ever in 2024.
05:27Very efficient result.
05:29Obviously, it's Coca-Cola.
05:31I think no one has a doubt on it.
05:35And it is also very efficient because it creates an emotional connection.
05:41And, yes, this one has used AI and is efficient.
05:45Let me share with you another example.
05:47I'm sure you guessed that it's not working, but let's see what it is.
05:52It is a quite well-known brand in Europe, eToro, but it's a German campaigner.
05:59Wer bin ich?
06:01Ich bin der, der 100 verschiedene Investoren sein kann.
06:03Alle auf einmal.
06:04Mit eToro kann ich Grace sein, eine globale Tech-Investorin.
06:09Oder Max, ein deutscher Experte für lokale Aktien.
06:13Ich kann Yam sein, ein passiver Dividendenjäger.
06:16Hey, ich kann damit sein.
06:18Dieser David.
06:19Mit eToro können Sie sich von Investoren weltweit inspirieren lassen.
06:23Ihr Wissen, Ihre Strategien, alles da, um es zu entdecken.
06:27Seien Sie der Investor, der Sie sein wollen.
06:29eToro.
06:32So, this is a campaign, or ad, a TV ad, fully done by AI.
06:38The script was written by AI.
06:41The video was produced with AI.
06:44The human are AI-human, you can see it.
06:47The voiceover, the music, everything was done by AI with a very small amount of human touch in it.
06:55I'm sure you agree that it's not the best ad you've seen.
06:58And I can tell you that it is in the worst 20% of the ads we have tested this
07:06year, or in 2024.
07:08So, obviously, it's not efficient.
07:11Clearly, AI has written a script based on the prompt where, first of all, the client probably wanted to say
07:19too much things.
07:19And AI will never say no.
07:24So, when you prompt to have a script, just keep in mind that AI will never say no to what
07:31you ask.
07:33AI will never say that it's not a good idea to say all that.
07:38AI will not also say that humans are not really human, are not enough, and there is no emotion.
07:47So, keep that in mind because what you should know also is that figures are 41%.
07:54We've done a study, and we already have 41% of consumers who are bothered by AI-generated ads already.
08:06Already 41%.
08:08So, be careful when you use it.
08:11When you use it, keep the key point of the emotion.
08:16Obviously, when we want to connect with humans, we need to connect with emotion.
08:21When you use emotion right, it creates four times more impact.
08:26And last, and it's very important to keep it also, we can clearly share with you that when we looked
08:33at all our figures at Kantar,
08:35creative, effective ads are creating four times more profit, four times more business.
08:43So, spending time on producing content with AI, great, but put human, put emotions.
08:53That's for content.
08:55Obviously, when we talk about attention, we talk about media.
09:01And more than that, most of it digital media nowadays.
09:05So, yes, again, for sure, thanks to AI, and we talked about it just before, you can target further, you
09:13can ultra-personalize at scale all your communication.
09:20You can even adapt the message of your communication to each person.
09:25That's great.
09:26And that, we already use it.
09:29But what we need to have in mind also is that tomorrow, your target will not be only humans.
09:39Tomorrow, and it started already in the US, your target will also clearly be the AI agent.
09:49I put some of them here.
09:52Obviously, everyone has created AI agents.
09:55So, Anthropic has one, Amazon has one, and of course, Operator from OpenAI.
10:01I'm sure you've seen all the video demo of Operator.
10:05Operator can order for you a restaurant, a taxi, travel, organize your holidays.
10:14The right question to ask yourself is how Operator will choose your taxi.
10:19Like now, you ask Operator to book you a taxi to go to the Champs-Elysées.
10:26How Operator will choose Uber more than Bolt or more G7, which is French?
10:31How?
10:32You don't know.
10:34You don't know how.
10:35Except if you ask, as human, if you ask them to choose G7.
10:42But if you don't ask it, it will choose for you and order it for you.
10:48I say E, sorry.
10:50E, I don't know.
10:52So, when you take the example, what is interesting to have in mind is that funnel.
10:59When you look at that funnel, yesterday, or let's say again a little bit today,
11:07human is the one who decides at every step of that funnel.
11:13Today, as you can see, a bit less human.
11:19And tomorrow, human will only decide the brand if he wants so.
11:26So, if he wants to have a specific brand, the human will ask to the agent.
11:33If not, he would not say anything.
11:37And, of course, I will be the one who will leave the experience of the ride with Uber.
11:43Okay?
11:44So, I will leave it.
11:45And I can put some comments on it, about it, and put a notation.
11:51And I can also share on the social media.
11:54So, what will stay human tomorrow will be awareness, the brand, and the experience.
12:03So, the key question now is, what will capture AI agents' attention?
12:12And that's a very new question.
12:15And we are all starting to work on it.
12:19And at Kantar, we also started, of course.
12:22But I'm going to share with you a study done by two Austrian men
12:28who has done a long study trying to understand
12:32what kind of elements that AI agents are looking at.
12:37And what they are showing, and it's a start, of course,
12:42is obviously AI agents prioritize structured data.
12:46Clearly, what they will look at is the price,
12:49the availability, the specification, the product details.
12:55They will look at that.
12:57So, very, very structured data.
13:00And so, just to make it simpler,
13:04in fact, when you talk about your project or your promotion
13:08or your offer with words,
13:11it's more efficient than when you put some words on an image.
13:16For the moment, the AI agent will understand further and better
13:22a text than an image.
13:25I'm saying for the moment.
13:27Of course, that will change and will probably move over time.
13:32But it's a key question.
13:33And when you start to understand that,
13:36you need to have that in mind and to work on it.
13:39And what is key for all the brands in the future
13:43is to understand that.
13:45One, we're going to have lots of different AI agents,
13:50not only one.
13:51I took the example of operator,
13:52but we're going to have lots of them.
13:55And depending on your sector,
13:57you're going to have different kind of agents.
13:59And you need to understand that how they select,
14:03because they're going to do it in a different way.
14:06I can already tell you that the one from Anthropic
14:09and OpenAI are totally different,
14:13and the way they select is different.
14:15Okay, so one, to which AI agent I am talking to?
14:21Exactly like if it is a consumer.
14:24And so how he's reacting,
14:26and how I should talk to him.
14:29I know it's strange.
14:30You're going to talk to machine.
14:33Not only to human, but we need to have that in mind.
14:36And you have to know that, yes, at Cantar,
14:39we are trying, not more trying,
14:41we are developing tools to help everyone
14:44to understand the profiles of each AI agent.
14:48And if your message,
14:50or the way you want to talk about your offer,
14:53your product, or your brand,
14:54will touch and will be selected by the AI agent.
14:58And that's really key.
15:01It started, I know that in Europe,
15:03we don't have access already to all the AI agents
15:07which are starting in the US,
15:09but it's going very, very quickly.
15:12And last but not least,
15:14I'm sure you understood it.
15:17In fact, the attention war
15:19will move to the recommendation war.
15:23Why?
15:25Because, I'm sure you get it.
15:27If I take the Uber example,
15:30I use Uber,
15:32I live the experience,
15:33I say that I'm really pleased,
15:35I share it with the social media.
15:37So, operator will see it.
15:39And the more he is going to have
15:41positive feedback from people,
15:45the more he is going to change it,
15:47to choose you.
15:49Clear?
15:49So, the war tomorrow
15:54will be the experience
15:55and the advocacy.
15:57In fact,
15:59what we already know
16:00is the key subject
16:02really for brands,
16:05besides to have a strong brand
16:07because you understand
16:08consumer can say to the agent
16:10that he wants that brand,
16:12is the experience
16:13and the advocacy.
16:14because that will be
16:16what will influence
16:17every AI agent.
16:21And it is already,
16:23to be honest,
16:23the case
16:24because when you look
16:25at the strengths of a brand
16:26and how a brand is measured,
16:29a brand is,
16:30most of it,
16:31built thanks to experience
16:32already today.
16:34So, it's going to be
16:35even worse tomorrow.
16:40Let me share with you
16:41just one example
16:42because,
16:43not because it's French,
16:44it's Sephora,
16:46but because it is
16:47a very good example
16:49of how you build
16:50brand experience
16:51using AI.
16:53It's Sephora
16:54and it's a few examples here.
16:56I'm going to take only
16:57the click and collect.
16:59Have you tested
17:00the click and collect
17:01at Sephora?
17:03What is really interesting
17:05is there are
17:06they have been
17:07the first one
17:07to do so.
17:08They are using AI
17:10to improve
17:11their human services.
17:14They are not using AI
17:15to replace humans.
17:17They are using AI
17:18to improve
17:19their human services.
17:21When you do click
17:22and collect
17:22at Sephora,
17:23you arrived,
17:24you think you're going
17:25to open that box
17:26and leave.
17:26Not at all.
17:28When you arrived,
17:29you opened the box
17:30and here you have
17:31a seller coming to you
17:32asking you
17:33if you're pleased with it,
17:34if you have any questions,
17:36if you want
17:36to have a demo
17:37of how to use
17:38that product
17:39or if you want
17:40to buy another product
17:41which goes with it,
17:43so you have
17:44real services.
17:45So even
17:47on something
17:48which is normally
17:49I win time
17:50and I don't use
17:51human at all,
17:52they are putting
17:53human.
17:54And the more
17:55we're going to bring
17:56human into
17:58the brand experience,
17:59the better.
18:00You can use AI
18:01but to serve better
18:02and to provide
18:03better human experience.
18:07So as a summary,
18:09we're going to move
18:10from attention
18:11to recommendation
18:15and there is
18:16three words
18:17to remember.
18:18If you manage
18:19a brand
18:19and if you're
18:21working in marketing
18:22and it started today
18:24but it's going
18:25to be even worse
18:25tomorrow,
18:26is brand,
18:28creativity
18:29and experience
18:30will be the three
18:32key of tomorrow's
18:35attention.
18:36Thank you very much.
Commentaires