Evan Shapiro (00:00)
I heard this unbelievable data point, which is for the last season of Stranger Things, Netflix created 1.5 million different versions of their trailer for YouTube and other social video. 1.5 million. There's no way that humans can scale that kind of editing thing. And this is what Aive is trying to solve.
Marion Ranchet (00:22)
Welcome to the Media Odyssey Podcast, that is Evan Shapiro.
Evan Shapiro (00:42)
And that is Marion Ranchet.
Marion Ranchet (00:44)
How are you, Evan?
Evan Shapiro (00:46)
I'm good, I'm good. I'm a little intimidated because there's going to be two French people on the pod today. So I have a petit un peu Français so I'll try to keep up with you two. But you know, we'll try to put it mostly in English for our American and British audience.
Marion Ranchet (01:03)
Of course. Yeah, bear with us. You're going to have a lot of French accents coming your way, but you guys love it. We know you love it.
Evan Shapiro (01:12)
I will say this is a good episode to check out on YouTube on video for a couple of different reasons. One, with all the French accents, we will be captioning it, but then two, we're going to be doing a really cool demonstration on screen that is best seen and not heard. Listen to it if you're in your car or you're on your jog, but then check it out on YouTube as well because we'll be showing the demo there. Let's get into some little bit of news. This is a big couple of news going on right now. Like it seems like the media universe continues to generate new headlines on a daily basis.
Did you see yesterday, OpenAI announced that they're shutting down Sora, which is kind of monstrummental news in the AI and media world. Did you see that fire across the transom yesterday?
Marion Ranchet (02:05)
Yeah, yeah, yeah. thought it absolutely... It's what, two years in? So was quite surprised, I have to say. I felt really bad for Disney on this one because the two struck a deal a few months back. And I think, you know, for Disney, the idea was very much to show that, you know, they were embracing AI, that they were innovative. Well, you know, until next time, we'll have to check back in. I'll be curious to hear Olivier's thoughts, our guest.
But we'll see more in a few minutes on that particular topic. But why is that? Is it a money thing? Is the product not where it should be? What's your take?
Evan Shapiro (02:44)
I think there's a number of different contributing factors here. Most notably, selling AI tools directly to mass consumers is not a business. There's not a real sound business model in there. And OpenAI is finding this out on a daily basis. We talked about this at the beginning of the year. I predicted that the AI bubble would burst this year, most notably because selling AI directly to consumers as its own product is not a real business model, you don't pay for email, you don't pay for search, they're wound inside advertising supported things, but also in other enterprise software.
And what they said was that we're gonna focus our business on enterprise software and not D2C software. So I think that's an admission that maybe, I don't, not particularly that I was right, but that their business model heretofore is a bit unsound.
And then the other part of this is I don't believe that there is a true market for moviemaking in AI. I think it's a great tool as a component of other things. So it's a great editing platform. It's helpful with graphics and sound and things like that. So it's one color on your palette. It is not the canvas. It is not the artist. And I think people are beginning to find that out. You we saw that epic battle between Tom Cruise and Brad Pitt created by some kind of AI. I don't know if it was Sora. I think it was another one.
But it upset people. It did not make people happy. People didn't go, ooh, cool. They said, this is creepy and weird. And then there's that Coca-Cola ad from last year, which was generated by Gen-AI. And it took, I think, 70,000 hours or some ridiculous number of human hours of prompting in order to get it to a mediocre version that, again, literally every audience member who saw it, hated. So I think this is the beginning of the end of the last era of Gen-AI. And I think we're gonna start to see it evolve into an arrow in your quiver, but not the bow.
Marion Ranchet (04:50)
Yeah, have to, a few things. Last week I was with RTL Ad Alliance and they had a trailer and one of the key stats was that 93% of social media videos were AI generated. Right? So there's one might argue that seeing Sora being shut down should actually go a bit wider than that because I do love, you know, to democratize, you know, video, put it in the hands of everyone, but right now we don't have any of the protection that needs to happen.
Like you just said it, there's potential copywriting infringements, fake news. It's becoming so, so easy to create good-ish quality content, video content. So in a way, I'm actually happy, I shouldn't say that, that they're gonna refocus on enterprise.
And especially, I think it goes very nicely with what I'm seeing with AI and what I'm expecting AI to do in our space, which is very much to be a B2B play and, to your point, focus on enterprise and be again, you know, and help a support.
Evan Shapiro (06:00)
Yeah, empowerment. Empowerment. Not replacing the artist, but helping the artist get better and faster. What I find interesting, so Disney was gonna put $1 billion into OpenAI, they're not gonna do that anymore. They saw it as a way to accelerate, actually, and I think lean into this new era of creatordom, which I thought was an interesting move.
We talked about that at the time, but then you saw Tyler Perry basically canceled his entire studio business after watching stuff being made by Sora, I think he looks a little silly at this point because I don't think we're gonna see, you know, Guillermo del Toro and Paul Thomas Anderson being replaced by Sora anytime soon.
But I think this gets into what the topic of this whole episode is, is how do you keep pace? If you're a studio or even just a creator, how, not just a creator, but if you're a creator, solo creator, if you are a creative in any form or fashion or a team of creatives right now, how do you scale social media video at this point?
Because we know this. We can't generate enough clips on a weekly basis to keep up with the pace of social video. And it's a real challenge. If you're at Disney or Warner Brothers, or if you're a Mr. Beast or Amelia Dimoldenburg, how do you satisfy the beast? How do you feed the beast of social video to the point where you can build the momentum you need for your projects.
And we have a guest on this week who, this is the entire enterprise. This is the whole concept of this business is to take the professionally produced content or creator originated content that you have and help you accelerate to the pace of social video today. Do you want to introduce them?
Marion Ranchet (07:49)
Yeah, absolutely. So we're going to welcome Olivier Reynaud. Olivier, welcome. I'll give a quick intro on who you are. You are the co-founder and CEO of Aive You're the first creative intelligence platform for video performance. You've been used by global brands, agencies, media companies. And like Evan said, it's all about maximizing the video creation process.
Historically, you're a successful entrepreneur because you co-founded Teads which is a global leader in video advertising. And so we're excited to have you today. Thank you for being with us us.
Olivier Reynaud (08:29)
Hi, Marion Hi, Evan. Thank you. I'm very happy to be here. So yeah, I'm from video with my partners. Also my co-founder, Rudy Lellouche we are both in the video industry since 20 years. So we know very well how video works. Previously, I co-founded teads.com. So we broadcast billions of video daily. So we know the pain point is not about creativity because with AI or without AI, we can produce video.
But the real pain point is to have an experience tailored for everyone. So large scale personalization for everyone in video. So that's very complex. So that's what we are solving.
Marion Ranchet (09:12)
And so how did you, when did you start Aive and again, what was that initial, you're talking about pain point. I love when technology actually solves this problem instead of creating more or inventing problems. So what was the genesis of Aive?
Olivier Reynaud (09:31)
Genesis was at teads and also my co-founder with Adyoulike the company he was driving. We broadcast billions of videos and we saw that when we personalize video for each social networks, for each media platform, it means not only the frame, it means the duration, it means personalization, the right content, the audio. We saw that these creative channels have strong impact in the performance performance, the experience.
So we saw it almost 10 years ago. And then after the acquisition of Teads I see that okay, distribution itself, we fix it. But now, make the creativity that we join how to perform with a video is something to do.
So how to adapt a video master for all experience? That's the topic. So at the beginning, we think with my co-founder Rudy, me, I have a creative vision and he have an AI and product vision. So together we have this vision that, okay, we will understand how the video is built in terms of creativity. So transform the video into data. We'll be speaking about it after. And after, thanks to this data, we can automate all the versioning, reframe, summarize, video optimization for social networks, localization, et cetera. So that's the genesis.
We have the vision 10 years ago and after we launched with Business Angels, we hired dozens of AI engineers and we developed the technology years after years.
Evan Shapiro (11:11)
So Bonjour Olivier. That's extent of my French right there. So let's go back in time just a hair because Teads is a massive platform. It really originated, I think, if I'm not mistaken, the idea of radical syndication of advertising. So an advertiser can come to Teads and then distribute its ads in contextual programming and content, whether it's print or elsewhere, all over the internet. So you empowered this multi-platform omnipresent distribution that became de rigueur, another French word, in advertising, video advertising around the planet Earth. Is that, am I getting that correct?
Olivier Reynaud (11:52)
Yeah, we saw the strong program that everybody is able to broadcast video campaigns for advertising, but also from a TV show and create clips. Everybody can do it. But the reality is nobody know exactly what we broadcast. So that means, for example, you have a one minute TV commercial from Coca-Cola. Okay, we adapted for an alpha minute for Society Works. Okay, we broadcast it.
We have the performance statistics that, okay, it was seen, not seen like this, okay. Is the message was seen? Is the agency was seen? Is Coca-Cola was seen? What the people, what the guys understand the video? Nobody knows it. So that's why having this data to understand what is inside the video and mix it when the guy see, for example, only 15 seconds, we know, we are able to say, okay, this is the beginning of the advertising, but he didn't see the brand, didn't see the message.
So right now this creative is beautiful, but is not adapted to the right people. So that's the problem. And that's why we created Aive.
Evan Shapiro (13:01)
So you created the opportunity, but at the same time there became this problem. So you enabled people to distribute their video everywhere, but until recently, or until let's say the last 10 years, there was a one size fits all creative brief. So coca-cola can make one, to your point, 60 second ad or whatever, and one creative too. One creative for the entire world where, so the opportunity became a problem, which is to your point, I'm not seeing the whole ad or I'm seeing the wrong version of the ad for who I am. So I'm, I'm a, know, Gen Z or I'm a Boomer and I'm seeing the same ad and that's not what radical personalization on the internet wants right now.
So you've created the solution to the problem you helped create to a certain extent, which is now we're going to take your creative and radically personalize it for the context in which the viewer is seeing it.
I just want to bring up one data point. was with Google last week, and I heard this unbelievable data point, which is, for the last season of Stranger Things, Netflix created 1.5 million different versions of their trailer for YouTube and other social video. 1.5 million. There's no way that humans can scale that kind of editing thing. And this is what Aive is trying to solve. Am I getting... Am I understanding this correctly?
Olivier Reynaud (14:25)
Yeah, that's it. So why they produce 1.5 million ads and formats? It's because 10 years ago, was, you say, one format that fit everyone. Now there is 20, 30 social networks with hundreds of formats. So you need to create all this format. Not always millions of formats, you have. But yes, when the broadcast is millions of formats, it's important to have to mix the creative scoring, the analysis of what is inside the video.
Okay, they produce all this video in a few months, but after, what is the performance? What is the impact? And can we link the creativity of the video and the performance? And after, thanks to this data, be able to create new versions. So creative A-B testing to better experience.
So they did a great part of the job, of the need by creating all this video. But the main thing is to bring the right creative to the right people. And depending on they consume it, they watch it, they experiment it, they adapt it, thanks to the AI at scale.
Marion Ranchet (15:34)
Yeah, that's fascinating. So who have been, you know, working with you guys? Who have started using Aive so far?
Olivier Reynaud (15:42)
So Aive it's a solution for enterprise. So there is brands, creative agencies and media. So we have many global brands with demanding companies, demanding needs like LVMH, Car companies at Stellantis, Match Group with Meetic, Procter & Gamble, Nestle. So huge company that from one video they need to adapt hundreds of formats in multiple countries need with all localization. And this, no creative want and love to do this adaptation, versioning, it's madness.
So this is brands, agencies, have Publicis Omnicom, IPG, Havas many global agencies. So we are scaling and deploying our technology inside the team. And in terms of media, we have Warner Bros, TF1, the number one TV company in Europe. And also we have some first GAFA we have Meta, is using Aive.
So we are scaling. So every industry that use video need Aive because video adaptation is for everyone.
Marion Ranchet (16:53)
Yeah, that's fascinating. So there's one thing is I love to talk about tech, but very often I love looking at tech. I think this speaks, you know, so many words. So I suggest we take a quick look and do you've done a demo for us, right?
Evan Shapiro (17:08)
Right, for our podcast. Talk about contextual. I think we did a demo, you and Jesse, Jesse's our editor and our producer. He's behind the scenes right now. We did a demo for the Media Odyssey podcast with this product, correct?
Olivier Reynaud (17:23)
Right, we did a demo based on the past episode of Media Odyssey.
Marion Ranchet (17:28)
And I think what's important to understand is that there's a lot of video editing tools. And so we're not going to be talking about what Aive does that a lot of other tools are doing, which is, you know, just like Adobe Premiere, et cetera. You can, you know, edit your videos. What we wanted to do, we took an already edited video and we were like, okay, how can we play with this to actually make it fit for more social media?
How can we make it fit for local audiences? And so we actually have a tiny bit of a surprise for you, Evan. I don't think you've seen this before.
Evan Shapiro
I have not.
Marion Ranchet
Yeah, we really took our hearts to use this to deliver on that versioning promise that Aive is pitching. So Olivier, we're going to show the demo right now. Do you want to speak to it a bit for us?
Olivier Reynaud (18:19)
Yeah, exactly. We'll see in the demo how it works. So just to summarize, we don't create the video here. We start from a video master. And with Aive we can do all the versioning. And here we will see exactly from a past episode how we can create a smart clip from specific parts that Aive takes the right parts, tell you this one is good, help you from an intelligent reframe, summarize, localization. Q uickly and adapted for all social networks. That's what we'll see and hear now.
Marion Ranchet (18:52)
Okay, so Olivier, this is Aive that's the platform. I'm seeing a lot of stuff in there. Do you want to tell us a bit more about what we can do in here?
Olivier Reynaud (19:03)
Yeah, Aive is a web solution, so it's not a software to download. Here you can see how we are on the Aive platform. So previously the video was uploaded on Aive and transformed into data. So it detects the people, the framing, the colors, the emotion, everything. And here, if you can see the video, we can see Evan speaking. We track him. So for the reformat, it's very great. And thanks to data, we detect the people, the tracks, the people, the face, the emotion.
So here we'll do a reframe, for example, a vertical format for the crop. So it's not just crop, it's tracking the right people. So because we have the data, here in this example, we select the right people, the right, so here we selected Evan, and after we click on generate the vertical format, and after it will do the adaptation automatically. And each shots, each chapters, each scene, it will add up and everything can be modified, optimized directly.
So here is the format, one of the hundreds of features. Here is the localization, we'll generate subtitles in multiple languages. So we want to understand more than 70 languages. We can just subtitle in a whole language.
Evan Shapiro (20:19)
For those listening really quickly, you'll see a vertical video oftentimes from a podcast, somebody will move out a frame. This prevents that to a certain extent. But then when on the toggle button on localization, there were all these languages for captioning. You don't have to think about captioning for all these different regions. It just automatically generates the captions in these different languages. Is that true?
Olivier Reynaud (20:45)
Yes, that's true. have the transcript in all language and we can translate the subtitle in all language. But after we can do a little more with the voice we see after. Yeah, we can customize everything like other solution with animation, your font, et cetera. We have all the parameters that are required for demanding agencies and brands. So we can do everything and edit everything like with a perfect pixel.
So here. We can see the demos that Evan is speaking, subtitle with animation. Okay, great. So after you can add your brand template. Yeah. Can add the logo just very easily. Have a, all the asset library and you can, if we have the video animation of the logo is a movie file. For example, you can add it like a TV commercial, but very easy to use. And after we say, okay, is my logo, subtitle the right place?
We have all the safe zone. or the grid, the magnetic grid, so you can add exactly the logo at the right place because it's different between an Instagram reel, a TikTok, a Snap, so we have all this information to have the perfect format for all experiments.
Marion Ranchet (21:58)
That's really cool. I have to say that, yeah, we've been doing this, you know, pretty much manually for the past year and a half and clearly.
Evan Shapiro (22:10)
Well we haven't but other people.
Marion Ranchet (22:12)
Well, you know, our team, it's a collective. Yeah, it's the team, right? I mean, we've been doing the pod for a year and a half. We're easily on all platforms when it comes to the pods, but all that, you know, discovery work that we need to do and, you know, that demand for social, Insta, LinkedIn, et cetera, it's been hard, right? So we've been able to do maybe a couple of formats. I don't think it's really perfect every time i see it especially on the subtitle
Evan Shapiro (22:43)
In fact, we've really decided to cut back on the number of formats because we just feel like we're not doing it well.
Marion Ranchet (22:50)
Yeah, we can't keep track, right? We can't keep track. So, you know, thank you so much for that demo. Maybe one thing that I want to show before we move on, because the versioning is very much the size, the frame, et cetera. But you've said it, it's also the languages. And so you can do subtitles. But so when we were doing this, we had a bit of fun.
Olivier Reynaud (23:11)
plus a creative score that shows you if the footage is adapted for each social networks and what you have to change. So that's really great because it's an AI video copilot. So if it's not adapted, you can see what you have to change or to add through the platform. So it's a copilot.
Marion Ranchet (23:29)
Yeah, for those who are listening to this, the score is 32%. It says it can be better. I'm thinking when I'm looking at that video, it's because we cut Evan's hair. We don't see Evan's hair creatively that's just not good, right?
Evan Shapiro (23:43)
That's right.
How do you give the people what they want?
Marion Ranchet (23:49)
Okay, cool. So just show us maybe the little thing we did for Evan. And then I want to talk a bit more about the data behind everything and that creative score you've just mentioned.
Evan Shapiro (24:02)
You're the McDonald's that just came out, or the Coca-Cola that came out using the IA. First of took 70,000 hours of solicitation by huge teams of people to create these really, really shitty ads that everyone hates and that are an absolute shame.
That was bonkers.
Marion Ranchet (24:22)
That was you in French Canadian, I think our friend Paul McGrath is gonna love this. So that was Evan in French Canadian.
Evan Shapiro (24:29)
That not just French, that's French Canadian?
Marion Ranchet (24:31)
That's French Canadian, it's not French French.
Evan Shapiro (24:34)
Get the fuck out of here. That is unbelievable.
Marion Ranchet
You liked it?
Evan Shapiro
Yeah, that was great. I suddenly can speak French, not just French, but Quebecian, I guess.
Marion Ranchet (24:45)
And so what's fascinating and Olivier we'll need to talk about that. So clearly we can hear it's still your voice. You know, so it's really you. It's not like there's a random voice going.
Evan Shapiro (24:54)
No! How does that happen? We want to move on to another thing. But seriously, Olivier, how does it take my voice and translate it not just into French, but into an obscure version of French that is just singular to the Quebec province of Canada?
Olivier Reynaud (25:20)
Yes, this is voice dubbing. So because we have all the data, all the transcription, all the information, we can after change your tone of, keep your tone of voice and adapt it in all language with all the controls. So that's part of some use case you can do.
Evan Shapiro (25:35)
That is absolutely bonkers. And by the way, that was also talking about what we were talking about at the beginning of this episode. Was about AI and that Coca-Cola ad too. That's really, that was well done Marion.
Marion Ranchet (25:46)
Yeah, well done, right? We did that on purpose. I think we did a good job. We wanted it to make sense for today.
Evan Shapiro (25:53)
That is crazy. You gotta send me that video.
Marion Ranchet (25:55)
Oh yeah, we're gonna send it to you and we have another version with me speaking Spanish and I'm hot. I mean, I'm so hot. I love it.
But so Olivier, I think I understand the problem, I understand the solution and what I'm seeing right now. And I'm not an expert, right? So the question is, what do you do differently than others? Are you the only one doing that or what's your secret sauce, right? What makes you not just a tool, but an entire creative and intelligent platform.
Olivier Reynaud (26:38)
You're right, we are not another AI tool. We are a creative, intelligent platform for video performance. What does it mean? Our huge technology is we transform the video into data. This is our technology, MGT, for multimodal generative technology. So this is a technology with more than 25 AI models. So each AI model transforms one creative topic aspect into data. We take the emotion, the framing, the type of object, detect the scene, the chapters.
So all the video, we transform it to data. We are doing video to data. So this technology is really huge. It's more than 20 million of dollars of research and development. And thanks to this data, when you own the data, we own, create the data, we can do everything. So we don't generate the first video, we can adapt. So all these hundreds of clicks of adaptation on the Premiere Pro, for example, or other software. With Aive it solved very quickly.
And also, thanks to data, we don't automate everything. We automate around 80% of all the production. So all the repetitive tasks, we automate it. And after, the human stays in the loop and through the Aive interface can do the final modification editing modification editing in order to have the result.
So on it, we can see if this technology has strong proof of performance. We see it because companies that take two months, three months to do hundreds of versioning, they do it in a few days only and with better quality at the end. So this this technology, MGT, that is a clear difference, but also with an entire value chain. So we understand the video, we transform it into data, we automate part of the production, the versioning. This is our creative, intelligent versioning.
And finally, we are launching the distribution. So thanks to all this data and the production, we'll be able in a few days to broadcast the right creative file to the right person on social networks and have the feedback on the performance. And then we'll able to create a better experience out of video. So data is everywhere on Aive platform.
Evan Shapiro (29:03)
I'm gonna bring up a point here, which is last week we talked about the Stream, the poll that Tubi and Harris Polling does. And there was this one really interesting data point that you brought up, which is people, younger consumers, all consumers don't mind advertising. They really don't. In fact, I think three quarters of them said, I don't mind advertising as long as it's contextual to my lifestyle, but also to the content that I'm watching.
That's great in a one or two channel world. You can kind of personalize that way by hand. But in a world where everybody's watching something different at different times on different devices, on different platforms, on different social video environments, to contextualize an ad for that many consumers on that many platforms and that many environments is impossible. It can't be done at human scale. This is the solution to that, this allows your messaging.
And for individual creators too, this allows your cut downs to match reels and shorts and snap, but also widescreen when it wants to be as well. It really is radical personalization that today's environment demands as opposed to, you know, let's pretend to make three versions, right?
Olivier Reynaud (30:24)
Yeah. your point something very interesting that there is a creative fatigue when you adapt, have the same video for all platforms. So I've had a clip for a TV show on all platforms is great, but many people who see the same ad, so there is a creative fatigue. So with a platform like us, you can do creative diversity. You can create stories on the same long format.
So that's why we recommend for agencies and that's what they are doing now, not producing a half-minute commercial, but a two, three, four-minute video. So with it, we can create an entire story. So thanks to this, there is no creative fatigue. And after, through AI, you can manage the experience through all platforms. So in this way, you are very consistent in terms of creativity on the global campaign and global experience for everyone. And so everybody can have the right experience, the right story fit for all formats, all montages.
Marion Ranchet (31:31)
And so what I love actually, which is new, so this is what you just said that you will be launching in the coming days, the extension of the platform. And that I love because what I'm seeing right now is everyone is selling tech to brands, agencies, media companies, studios, you name it. And it's all those bits and pieces that are supposed to be working together, but are not. So the fact that you can do a standalone video editing job, then do the versioning, then distribute directly, right?
So instead of having a manual moment where, you know, all of those things you've created, you need to push them individually, et cetera, on XYZ platform. Being able to do that within your platform, I think that's a super, super smart move because you said at the top that we fixed distribution, but distribution is still complicated. So you do the creation, you do the distribution. And what I love is the analytics parts, because that example you gave Evan at the top, right, with a million plus of videos for their latest series, I can remember which one you mentioned. But yeah, the fact that you can then understand what works, what doesn't, and recreate on the fly additional versions that actually matched that for a particular target audience, a particular country.
I think that is fascinating because that's the one thing, right? It is 24 hours. You need to push as much content as you can. But when that next 24 hours starts, are you just going to go and repeat? Ideally, you understand what happened the next 24 hours and you tweak and adapt and then you keep at it. Right, Olivier?
Evan Shapiro (33:18)
It takes A-B testing to that, you know, it allows you to do A-B-C-D-E-F-G testing as opposed to...
Marion Ranchet
Exactly.
Olivier Reynaud (33:25)
There is this, but also in terms of data management and security, because we are all in one platform and we cover the entire value chain of video analysis, production, distribution. For the money company, it's mandatory. If you look about Disney, if you look about MrBeast, when they produce a video, they won't export it and upload it on a SaaS platform. Videos that cost millions of dollars, they won't upload it everywhere.
They have to have something very secure, all-in-one platform. And our technology is proprietary. It's not trained on OpenAI or Amazon or Google. So when a company send, for example, the future video of the next smartphone, the Future car or the next movie with hundreds of millions of dollars of cost of production, it's secure, it's safe. It's a SOC 2 platform. So it's totally waterproof in terms of media data. It stays here. It doesn't train outside. All the data stays here.
And how it works, I'm not an engineer to tell you more precisely. This is the technology you created. This is meta-learning. Meta-learning, it's not one fundamental model that generates video or content. Because we got many, many AI models that we own together, we generate the data. So we don't need to train a model like all the other GAFA companies. And to be true, LLM, for example, just 5 % of all the technology. So all the technology is in-house and don't train outside and stay safe.
So that's very important for all the brand agencies and media companies for the content.
Evan Shapiro (35:10)
Yeah, so and and Olivier, feel free to not answer this question. But my my curiosity is, is this only accessible for a company the size of Disney or Disco bros or Meta or can an individual team like ourselves or creators is this is there a scalable version of this that allows the individual creator to use it as well?
Olivier Reynaud (35:33)
So yeah, the solution is right now available for enterprise or brand agencies with this kind of size like a Publicis, Disney, et cetera, but also available for medium team. So medium size of agency, medium size of brands or maison or media company, it works. But
Evan Shapiro (35:54)
It's not just a billion dollar company? It can be for, you know.
Olivier Reynaud (35:57)
But it's not today a solution for individual creator at $20 per month, maybe one day, but not right now. I focus it for very complex needs. So you upload a two-hour movie. Okay, we know we transform everything. And for this two-hour content of Disney or MrBeast, you can create hundreds of clipping adapted for all social networks, but not enough in a few days and it will be better done than it was done by humans. And that's what we see in all the use cases we have with a huge campaign.
Evan Shapiro (36:38)
Well, and it's a number of versions that allows. That's where the difference between, know, I mean, human editors are good, but they can't just do a thousand different versions in an hour.
Marion Ranchet (36:49)
Despite being able to do it, there's the question, perhaps they can be doing something different with their time, something more valuable. And so the question, I think we've heard you say that you're trying to kill that 80% of repetitive tasks and then you want creatives to focus on that last mile, that last 20%.
I have a feeling, but I'm interested to have your thought. It feels like you're not seeing AI as a threat to creativity, but really more as a way of augmenting creators, artists, et cetera. Is that the case?
Olivier Reynaud (37:31)
That's the case. This is augmented creativity. To give you an example, we have one of our clients, Meetic for the entire Europe for Match group. They created a campaign on the entire quarter with almost 300 of campaign. So that's not what just one try. And so they have to adapt all this format for all social networks with Aive.
And what is the result? This saw that they save almost 80% of the cost of the production. Also, the time of production with Aive of the versioning with Aive they can produce four, five more content than done with Premier Pro. So the go-to market is really reduced from two months to a few days and weeks. And finally, they broadcast and they do campaigns on Facebook paid and Instagram paid and see the results and all the formats adapted with Aive have a performance uplift of 50%.
So it's 4x faster and it's performing better. So we succeed to match creativity and the performance together. So to answer finally to what Meetic can do with this time saved and the cost. Now they can invest more in the media so they can do more advertising.
Also, they tell us that they are able to invest more in the production, create more movies, more masters with creators, with agencies, or with Gen AI And also, they are able to give more trainings for their teams. So nothing is lost. With this time and the cost saved with our technology, they can create, finally, better experience and better performance for all the companies.
Marion Ranchet (39:27)
Fascinating. Okay, so at the top of the episode, we had a quick discussion about news and there's always something fascinating happening in the industry. And we just heard that OpenAI is actually closing down Sora, at least the B2C component of Sora after two years in, which means that they're gonna stop their deal with the likes of Disney. How are you seeing this move? Is this the right move? What does that tell you?
Olivier Reynaud (39:56)
Generated video right now is a commodity. Many companies do it. There is Google, Kling, Runway, et cetera. All models are really great. Every month it's improving. So now it's not a deal. The deal for OpenAI is to concentrate on the core business for enterprise like Anthropic. So I think it's a good move because for creators they are not lost. They have a lot of solution for generation.
Maybe that was the beginning, the birth of a new social networks. And we saw also there was an AI slop There was a lot of video generated at Sora that was everywhere on Instagram, TikTok. It was an AI slop moment. So I think it's a good thing for OpenAI that we focus more. And so it gives opportunity for other company to focus on the true value.
Otherwise they focus only on video generation. That's it. So for video generation, it will improve in the month in the year to continue. But the topic about scale adaptation versioning at scale and the experience is another topic. And on our side, just so we saw, because we have some agencies that use Aive are agencies that only generate video. So videos that generate can only video with a company like Runway, et cetera, they can use Aive for all the adaptations. So that's not a big problem for the creative industries. And so for Sora, for OpenAI, will focus more.
Marion Ranchet (41:37)
That makes sense. Right. So there's a lot of things happening for you guys in the next few weeks. You are going to be at NAB in Vegas. There's at the same time, the Adobe Summit. We're going to show for those of you who are watching this, or you can go check the show notes, but the team at Aive is giving us a code so that you can actually have a free pass to go on the show floor. You guys will be demoing. You'll have a booth?
Olivier Reynaud (42:02)
Yeah, have a booth at NAB show and Adobe Summit also. At the NAB show, we have some pitch also on many days. So you'll be able to see Aive live to meet the team because we come with many people. So let's meet us at the both events. And also we have some free pass to give for NAB.
Marion Ranchet (42:23)
Nice. Okay, cool guys. That was awesome having you on the pod. Olivier, thank you so much. It was really cool. Very cool. Very funny as well. As you could see, I really love my French Canadian co-host, Evan Shapiro.
Evan Shapiro (42:40)
Shapiro?
Marion Ranchet (42:42)
Evan Shapiro speaking Canadian, French Canadian. Thank you for being here. Thank you everyone. This was another episode of the Media Odyssey podcast. That was Evan Shapiro.
Evan Shapiro (42:54)
That was Marion Ranchet and we will see you again next week.
