AI in Marketing: Generative vs. Regenerative AI with Cory Treffiletti (Ep. 2)

Adam Boettiger: 00:00:00

Today, we're thrilled to welcome Cory Treffiletti, chief marketing officer at Rembrandt with over two decades of experience in digital marketing. Corey is a seasoned chief marketing officer who's led efforts at both startups and Fortune 300 companies.

Before joining Rembrandt, Cory held leadership roles at firms such as Blue Kai, which was acquired by Oracle, and Voicea, acquired by Cisco.

At Rembrandt, Cory focuses on using AI to drive digital growth for B two B and B two C brands, helping companies unlock new value through innovative media solutions. He's also a published author, contributing regularly to industry discussions, including his original digital newsletter and columns in media post.

We're excited to dive into his expertise today and here's insights on how AI is reshaping the marketing landscape we're going to talk about generative AI and regenerative AI today, and most folks know who are watching or listening to this what generative AI is chat, GPT, Claude, Gemini, Copilot, all of them. Everybody's experimenting with it and it seems to be, let's just say the hype, really hard to cut through the noise.

Let's have a conversation between you and I. I'm a chief marketing officer myself with decades of experience, just like you. Let's talk about marketing and generative AI for the time being.

So how generative AI automates content creation from text to video, how marketers are using it to enhance efficiency and experiment with personalized campaigns, and maybe touch on one or two real world examples of generative AI's impact on marketing teams. So what are your thoughts on all the hype?

Cory Treffiletti: 00:01:46

I think that the hype is natural and it's normal. I don't think it's overhyped. I also don't think it's under hyped. I think it's about right right now.

Why I say that is because when it first came out that everybody could start using these tools, which is really about two years ago now, it seemed like an overnight thing. I know it was probably in development for years before that, but it seemed like overnight.

Two years ago these became available and on everybody's doorstep for years. These days, I already feel like the hype died down a little bit because there are questions about how to use it ethically.

There are questions about how businesses are going to be permitting it, being used inside their walls or on their tools themselves.

But fundamentally, I think people like you and I, most marketers are using these tools on a day to day basis already, and it's in the role of a co writing partner. It's in the role of a tool for ideation or I collaboration to come up with ideas. It's in a way to edit things like editing videos, editing content.

There's a lot of really basic applications where people are using this right now to speed up their processes for marketing. And fundamentally, marketing is a copy is a combination of things where you get to be creative and you get to be analytical.

AI can speed up both of those processes and I think that's where people are using it. And it hasn't replaced anything. In my eyes, it's something that is allowing you to open up portions of your day in a way that you couldn't before.

But I don't see it replacing anybody's specific role yet. Maybe some of the responsibilities.

Adam Boettiger: 00:03:22

Yeah, I agree.

I think a lot of people are saying that you're not going to be replaced by AI, you're going to be replaced by somebody who knows how to drive AI, and I firmly believe that. So everybody should be taking a deep dive into it.

I myself use generative AI on a daily, almost hourly basis, and it's not to just create stuff and then claim it as my own. I use it a lot for ideation, I use it a lot for analysis. I have a lot of documents that I'll put into it after a course.

I scrub it for Pii and anything confidential, and I'll ask it to analyze this and give me ten ideas on that. But then I take it and I use my own creativity and I change it and I make it my own.

And I think that's really, I think that's a good use of generative AI.

I think what people need to be careful of is just using it to rush out at the last minute, that piece of copy, or that, the report that you have and then claiming it as your own and sending it off without checking it. Tons of incidences of hallucinations.

And there was that incident, you may remember, I think, last year, where a lawyer was preparing a brief for a court filing and he had chat GPT create it and didn't check it and didn't recheck it and follow up on it. And it turns out that Chachi Pt hallucinated court cases that never happened, incited them in the briefing.

Cory Treffiletti: 00:04:51

All these tools, you still require oversight on all of them.

Adam Boettiger: 00:04:55

Yeah.

Cory Treffiletti: 00:04:55

And if you don't, to your point, you're going to find hallucinations baked in that CrE, incorrect statements, whatever. The other thing is, it's really easy to figure out when somebody used a chat CPT or some other Ji tool to write something.

Yeah, it doesn't have a character, it doesn't have soul, it doesn't have kithiness. It doesn't. You can ask it to write that way, but it just doesn't work.

There is a humanity intangible component of a piece of writing that only humans can do, and maybe eventually that won't be the case. But for right now, it absolutely is very clear.

Adam Boettiger: 00:05:31

I totally agree with you. I think there's one word, I forget what it is. I should have probably looked it up before we got on the air.

But there's one word that it keeps wanting to inject. It's some kind of adjective that I see it in every single thing that you would normally never use in conversation. So that's funny. But let's talk.

Let's touch on what you said earlier about businesses specifically, maybe marketing teams or almost any business wanting to implement generative AI, or even being really worried about it, because the reality is that your employees are probably already using it, and so you should have some kind of, I don't know what, query, oversight, plan, strategy, something to make sure that they understand what should and should not be shared with it, and also a way to technically keep the generative AI from training on your company's information. What are your thoughts on that?

Cory Treffiletti: 00:06:28

A lot of companies have been afraid, and you said it earlier already, they're afraid about personal information or proprietary information being input into those systems and then used to create whatever output is necessary from a business perspective or a legal perspective or marketing perspective. And once you put something into an LLM, it is available to be used in any other way, shape or form that somebody else is going to find.

Now, the reality is it's really hard to go find any kind of specific information.

Like, I can't go to a chat, GPT or copilot and say, what have these five banks or what has these five agencies input into your system that they wouldn't want me to see? I can't ask it that question and have it directly identify and extract that pieces of information.

Which kind of goes back to the overall challenge of these tools, is that when you give them a specific constraint for real pieces of information, a real world object or real world scenario, they're not capable of doing that yet. The fact is that companies have almost an unnecessary fear of the risk associated with these platforms.

But to be honest, it's good that they do, because we don't want to inundate these systems with a bunch of information that's proprietary and that is PII, and that can get used in the incorrect fashion later.

I think if you look at the ways these platforms should be used and the compliance and the governance and the regulatory elements that most businesses are adopting, to be wary of putting too much in, I think it's good because on one side, there's not going to be this just wealth of personal information that's available. But on the flip side, too, these platforms are being trained on whatever data it ingests.

And unfortunately, it's ingesting a lot of really bad information and bad data. Just the fact that these AI models are being trained on the Internet means they're being trained on a lot of, oh, my gosh, not factual information.

Adam Boettiger: 00:08:23

Yeah.

When somebody asked me, and I don't know if it was maybe my father, who's in his eighties, or my mother, who's also in the eighties, both of whom are a little bit tech savvy but not too much, and they asked me about what an LLM is and where the AI gets its information, and is it? The question that they ask a lot is, how do we know that it's factual? How do we know that it's true? And my answer is always that we don't.

That's why we have to check it.

That's why they had that little disclaimer below the chat box on Gemini and on Copilot and Claude and chat GPT, that it makes mistakes, because it does very rarely.

Cory Treffiletti: 00:09:04

It's even better than the same mistakes. The concept of garbage in, garbage out.

Adam Boettiger: 00:09:08

Yeah.

Cory Treffiletti: 00:09:09

It's not that it makes a mistake, it's just that what was, what it determined to be real is not.

Adam Boettiger: 00:09:15

Yeah. One of my favorite places to get information is Reddit. I'm a huge fan, been on there for years.

And I think, you know, I believe that there's traditional education and then there's all kinds of channels where you can get education from people and books and engaging with, with other people online or podcasts like this on YouTube videos.

And one of the things that I heard, and I know Reddit's great for information, it's great for finding answers and stuff, but when Google did that deal for, I don't know if it's $150 million or something this year, where they bought the rights to train on all of the data in Reddit, I was like, because it's not all accurate, it's just people's opinions who may or may not be experts.

And the same thing with it, I think it's safe to assume that it has ingested not just all of the web, but since we have archive.org, comma, all of the web that used to be the web. And if anybody can put up a website and anybody can say anything about anything, I think it's a fair question to ask.

How do we know that what is being spit out by these things is factual?

Cory Treffiletti: 00:10:22

Yeah, I want to see the LLM that's been trained on all the data from geocities, Lycos and Infoseek and see how much that has changed.

Adam Boettiger: 00:10:30

Yeah, or the good old bbSs back in the day. That brings up an interesting thing.

I think there's going to be value in having the huge LLMs, but there's also going to be tremendous value in businesses having their own private repository that they upload documents to and customer data and behavior that they can analyze, especially for marketers, especially for customer service folks.

Other than generating copy, what are some real world examples that you can share with us about how either you use I could talk about how I use it for generative AI, and we can talk about certain ones are better at things than others are. I use Claude for writing, for example, and chat GPT for the most part. Everything else.

Let's just talk about a couple and then we'll segue into the second part of this, which is talking about regenerative. Aih, what do you use? So which ones do you use out of chat? Gt Claude Gemini Copilot I actually primarily use copilot.

Cory Treffiletti: 00:11:34

I've been a paid subscriber to Microsoft Office for many years. I'm probably one of the few remaining people who still uses it. I like all the typical tools in there.

So I have Copilot and I like the way that it works. I think it's pretty accurate.

One of the things I've really enjoyed with with Copilot is how it cites its sources for a lot of the information and builds an index of sources below the content. I really enjoy that because it allows me to go check up on what is extracting from me.

To your point, I do use it for not writing because I only pride myself on the fact that I write everything myself, whether it's personal or work or my blogs and my columns or whatever. But I will come to the platform to help me generate ideas. I might ask it an opinion.

I might ask it to go look up some sources for me and help me understand differing sides of a point of view that maybe I don't fully understand. So for me, I like copilot for that perspective. I also use it for image generation. I also like PI.

I think it's been a little bit glitchy lately, to be honest with you. But I put PI on my phone and I'll just put my phone down and talk to it as if I've got somebody in the room because I be typing all day.

And so I will sit and have a full conversation with PI just to get an idea once again of a back and forth in a conversation that helped me generate an idea that then I can go on and I can write about. I do use both of these things typically to try and understand.

I think I mentioned already the different point of view than what I have because I tend to have an opinion on a lot of things, which is good and it's bad.

But I'm also very open to hearing the other side of everything, whether it be politics to the state of different types of technology and how to use them.

Adam Boettiger: 00:13:15

That's a great use of it, to tell them to take on a contrariety viewpoint.

Cory Treffiletti: 00:13:19

Yeah. Just so I can understand it. We're going to agree with it. But I like to know what somebody else might think and that I didn't.

I don't want to be in my own echo chamber all the time.

Adam Boettiger: 00:13:28

Yeah, that's good. So one of the things that came to mind when you were speaking about PI is, man, this is, I don't know if it's just me, but this is like a fire.

This, if the Internet back in the day was a fire hose of options and choices. I don't even know how to compare this because this is, I had totally forgotten that PI existed.

And I'm sure that there are people who are watching or listening to this that are just overwhelmed by these choices. Back in the day, we had the browser wars with Netscape navigator and Internet Explorer, and now I'm dating myself, of course.

But you had, it was like four choices, but now it's like, how many people can I give $20 a month to? It's. And then remember that they exist so I can. It's going to be such a, I know it's going to, I know it's going to play out.

I know that the player, probably 98% of the players are going to leave and there's going to be, we're going to be left with some major ones, probably Google, Microsoft, the big players. OpenAI. Another one that we haven't talked about yet is perplexity. And I find that's a really interesting engine to use.

Cory Treffiletti: 00:14:39

And I had this conversation happen to me twice in the last three weeks, there is an entire portion of the audience who no longer uses Google. They only go to perplexity to ask questions. They tend to skew younger.

Google is not a source of truth or if any information really, for kids these days, anybody under age 18, they don't use it.

Adam Boettiger: 00:14:59

Yep.

eemed like it just blew up in: 2022

But I think that I saw it come out and the minute that I saw that, I could just type something in and it would give me a detailed response and then expand on it if I wanted it to. I thought, what's going to happen to Google advertising? What's going to happen when people don't visit websites anymore, when they just go?

Because that's really the majority of people, they go and they search the old way of doing things.

You go to Google, you search for something that you want to find, you're presented with a list of four people that paid to be there and then six that didn't, but are just ranked organically because Google thinks that's the right place to get the answer. And then you have to go to the site and then hopefully you'll find it by digging down.

revenue models in February in: 2022

Cory Treffiletti: 00:16:17

That I said, I expect web traffic to go down by 50% in the next ten years just because there's no reason to go anywhere. The joke that I pulled, and I actually used it in a media post event, which is worse.

If you were to go murder someone, where's the best place to hide the body? It's on the second page of Google search results because no one goes there and no one should ever see it.

Adam Boettiger: 00:16:37

Exactly.

Cory Treffiletti: 00:16:38

Yeah, that's going to be an absolute fact.

I think websites become knowledge bases, which is just basically being scraped for data and information to be able to extract back into a Ji search result. But it's really interesting is that you touched on it, too.

This is a real interesting time for the government to be going after an antitrust case on Google.

When the people are skipping the ads on their video platforms, they're not going to go to the search anymore because they're using Genai for perplexity and other tools. It literally coincides with the very first time Google has a threat on its business.

It's interesting DOJ can actually go after Google and try to break it up, or it could just wait this whole thing out and potentially Google will see a decline in its overall ad revenue anyways. Not to mention the big beast in the room, which is Amazon.

Even in the last 24 hours, I think they were talking about, Amazon did a big presentation somewhere, and they're coming after all advertising. They're not just going after retail media networks, they're going after all advertising.

And that's if you're Google, you got a war on three fronts now.

Adam Boettiger: 00:17:40

Yeah, definitely. And like I said, I think the major players are going to shadow everybody else and they're going to force them out of the market.

So which, going back to your preference, and by the way, I don't know if you knew this, but I, you said that you used PI, and PI is a great app on your phone that you can use for talking to, but chat, GPT also has an app and I have conversations with that. The only thing, it's just, it's almost there, but it's not quite there. Cause it interrupts me a lot when I'm talking.

Maybe it needs to give me a few more seconds. I tried telling it that and it apologized to me profusely, but.

Cory Treffiletti: 00:18:14

Correct. At least we're not sitting talking to ourselves.

Adam Boettiger: 00:18:17

At least we're not talking to ourselves yet. Yeah.

So getting back to originally, we had talked about some real world examples of generative AI and how marketers are using this copy ideation strategy planning.

As I just as a full blown grown old adult, I went back to get my MBA this year and there was a lot of talk about AI, as you can imagine in school and in the courses and stuff, the difference between critical thought and AI and how it's a huge struggle among educational institutions right now because they're always, I hate to say this, but they are very slow to adapt new technologies and to embrace them.

And especially with something like AI, which is going to be such a disruptor on so many different levels in education and teaching, I think that they have, they have a right to be concerned about that and they're going to, I think it's going to completely disrupt education. And yet at the same time it is. And I tell my parents this a lot, it is perhaps the greatest advancement in learning opportunity in our lifetimes.

And when you think about that.

It can teach you the same things that a person can teach you, but it can do more than that because it has multiple opinions and all of this data, and it can do it faster. You can ask it to, if you're new to marketing, teach me how to develop a marketing plan and explain it to me like I'm 16 years old.

There's a lot of wonderful things. And then I find people I talk to who are shying away from this, who are not luddites, but pretty close to it, who are just, this is Skynet.

It's gonna be. It's gonna be sentient, and we need to just stick our heads in the sand and not do it. I think that's not an answer, but.

Cory Treffiletti: 00:20:12

It'S too easy to just sit on fear. Yeah, it's too easy to say that. It's going to make people lazy.

You can go back and say that the calculator was something people worried about, so they weren't exactly math, and that Excel was gonna be something they didn't like anymore.

There's always a portion with an audience that's going to say, this is something, and it's a negative influence on our ability to be educated and learn, et cetera. There's always a group that's going to say, I'm going to take this and run with it.

The digitization of everything scares some as opportunity for others, but it's inevitable.

Adam Boettiger: 00:20:46

Speaking of digitization, how I think people who already have ADHD, if they didn't have it before AI came along, everybody's got it now. And I remember when the Internet first came along and the web, and it was just like, holy crap, this is like a fire hose of information coming at us.

How are we going to get any work done when there's so many fun things to explore and see and learn? And so how do you. Everybody has prompts that they use and reuse and prompt libraries. How do you organize that on your end?

Do you have marks for all of the things that you pay $20 a month for, and text files that you organize by prompt, or you just recreate every time?

Cory Treffiletti: 00:21:30

Good question is, I'm sure how I.

Adam Boettiger: 00:21:34

Answered, to be honest with you, because.

Cory Treffiletti: 00:21:35

I purposefully keep my sphere of paid influence small. I don't want to pay for a whole bunch of stuff. I only pay for a couple of things, and I try to get really good at using them.

I will set aside a portion every month of my time or my day to go mess around with some stuff, or if I see something that I want to go learn about. That's usually like my late night excursion, because I'll go and play with it and figure out if it's something that has value to me or not.

I also tend to like, and this is different, I tend to like just talking to people like we're doing right now, go have coffee with people or whatever, and ask them what they're doing and what they see and what's interesting to them, because I find a 15 minutes conversation with somebody, I'm going to get more value than if I had just tried to scour around and discover everything myself.

Adam Boettiger: 00:22:23

Yep. Yeah, I totally agree with that.

Cory Treffiletti: 00:22:26

Yeah. I like to take cues and tips for people who are smarter than me, and it's not hard.

There's a lot of people I can find were smarter than me, so it's just a matter of me getting out of my room.

Adam Boettiger: 00:22:35

Yeah, I think part of, and you and I probably have, between us, 50 years of marketing experience, and I think that part of being a good leader is recognizing that you don't know it all and tapping into the talent around you, which includes the 21 year olds and the 20 somethings who are. Everybody's an expert at something.

And I discovered something early on, which was, if you humble yourself and realize that it's impossible for you to be an expert at everything, but you can learn something from everybody you come across. I try to do that. Even like a waiter at a restaurant, you can learn something from them.

And it's important to expose yourself to those things and as a leader, to recognize that it's not just about giving orders and telling people what to do. It's a collaborative effort.

And I think that's more important when AI came along, because there are going to be people, everybody, all of a sudden, AI came along. Everybody's an expert at it. Take my course for $5,000 and I get it.

But there's no way everybody can be an expert at something that changes, like every Tuesday afternoon. It's just not possible. We have to all learn this together, and we have to all go on the journey together. Let's talk about regenerative AI.

You've come up with something at Rembrandt called regenerative a.

We see on movies where people pay for product placements and they ship Apple computers and MacBook pros into the scenes of movies and tv shows and pay for that. And then sometime down the road a year from now, they look at the results. So can you explain to us what regenerative AI is?

And what the key difference is.

Cory Treffiletti: 00:24:13

Yeah. What I'll do is I'll explain it in the way that it's explained to me by our technology team, because I'm not the technologist.

I know enough to be dangerous.

Adam Boettiger: 00:24:21

Yeah, good. I'm not either. That's not what this podcast is about. We want just non technical stuff.

Cory Treffiletti: 00:24:28

Yeah.

So let me actually, this way, if you were to basically go to a generative AI platform and say that I want to create a video of a black cat walking through a living room that has white walls and a blue couch, and I want it to look like it's the middle of the afternoon, so I want sun shining in the window. It can do that. And you'll get back a video that looks like that. Exactly.

ing through my living room at: 02:00

What we have done with regenerative AI is basically enabled you to take two objects. One can be the cat, one could be the living room.

And if you've got a video of both, we can actually merge them together in a way that does look completely realistic, as if the cat was in the room when it was originally filmed, not that it was put in afterwards.

Because what regenerative AI, and specifically called regenerative fusion, what it does is it takes a video, which to you and I is a two dimensional piece of video content, and it renders it as a three dimensional landscape, as if it was actually there when it was being filmed.

Then it takes that other object of the black cat and it inserts it in, as a 3d model into a 3d reconstruction of that room, and then basically merges it back together into a two dimensional fashion so that cat could walk behind the couch and jump up on the couch and be in front of the window, and your hand will go in front of it and it looks completely natural and realistic. And that's what regenerative fusion does. It enables us to put brands into videos that other people created.

But the brand looks like it was originally there when it was created. And it's not an overlay that sits in front of everything else because that would be easy. And it's not a flat object because that would be easy.

It has to be a three dimensionally rendered object that reflects lighting and has the same look and feel as if it was there when it was shot.

Adam Boettiger: 00:26:54

Interesting. So, like you said, you and I know enough about it to be dangerous.

We know how to drive it, but we don't need to know how it works necessarily from a non technical standpoint. Say I am an Apple computer and I want to.

Historically speaking, I've been doing product placements, shipping my laptops to high ranking series and to get them in. It's interesting, I read something that Apple has a policy where they can never.

Their products can never be placed and used by bad guys in series or movies. I thought that was interesting.

But say your Apple computer and you're used to just doing this physical thing and seeing the results after the series comes out maybe a year later. So am I right in saying that somebody could take AI and you could put that into media that's already been created, or a historical library?

And what is it on? Is it on all the social networks? Is it on YouTube? Where's this placement? Where are these placements and how targeted are they?

Cory Treffiletti: 00:27:58

So I will one up what you just said in a good way. So it can be placed into a library of historical video content. It can be placed into content that's about to be published.

It can be targeted so that different people can see it in different libraries of content, and it can be done in live video. So all these things are doable, whether they're doable today or they're doable in the next few months, that's obviously a different conversation.

But all of them are doable because you can monetize the content the way that doesn't interrupt the content by doing this. And there are different extremes of the creative.

So some of it can be very natural, some of it can have animations to call attention to it, but it's all very doable right now. And it's being done for us. We're doing it on creator content on YouTube and TikTok and other platforms right now.

But we're also in conversations to sell the software to a lot of other people. So they can just do it themselves on whatever content they have. Whatever video content they have.

It could be major studio stuff that you see on the major streaming platforms. It could be in mobile platforms, it could be in smaller niche publishers. It can be basically anywhere that there's video.

You can monetize that content upstream of the ad tech stack.

Adam Boettiger: 00:29:04

Regenerative AI seems like a pretty broad term. What's the product called or what do you have there?

Cory Treffiletti: 00:29:12

We have both of which we have copyrighted packets on our spatially aware AI, which is an AI that can recognize the three dimensionality of a two dimensional space immediately in real time.

And then the regenerative fusion, which is the application of that 3d rendering into a two dimensional video, which has now been rendered as a three dimensional landscape and then reconstructed as a two dimensional video.

So those two things are the core components of the technology we have, in addition to a couple of other things we have around brand safety and creator selection, and some other things that we've got to work with too.

Adam Boettiger: 00:29:46

You guys are a startup, right?

Cory Treffiletti: 00:29:48

Company's been around for about two years. We've been actively selling it for about a year and a half in the market.

Adam Boettiger: 00:29:53

Let's talk about, I think you mentioned during the break that you had called it insane media.

Cory Treffiletti: 00:29:59

Yeah.

Adam Boettiger: 00:30:00

Okay.

Let's talk about something that's a challenge to almost all advertisers, especially video advertisers and marketers, which is reaching the unreachable audiences. It tells who the unreachable audiences are and why it's a challenge to brands to reach them.

Cory Treffiletti: 00:30:20

It's basically you and me and a whole lot of other people.

If you've ever paid for a premium service that doesn't have ads on it, which I think I pay for three, if you've ever put an ad blocker on your computer, which I don't currently have one, but I've used them in past, and if you've ever had that urge to skip an interrupted video ad whenever you possibly can, which I think is almost every person, then you're basically part of that unreachable audience. You're the audience who has found ways around advertising, especially in a video environment, but also in other places too.

And you are typically enough good income to be able to pay for premium services. So you're a little bit higher income families typically. You're probably pretty well educated, you're probably pretty prone to technology.

You're much more an early adopter slash mainstream adopter than you are a laggard. These are audiences that most brands want to talk to.

So the fact of the matter is that the audience that most brands want is literally trying everything they can to avoid the ads from those brands.

Adam Boettiger: 00:31:23

Earlier you mentioned that there was a case study that you could tell us about of insane media driving audience engagement, and then also perhaps speaking to how the audience provides feedback on it, they can then iterate on.

Cory Treffiletti: 00:31:38

Yeah, not to do like a full on case study, because I don't want. We've run dozens and dozens of brands just in the first half of this year.

And if you look at the numbers that we've seen, on average we see about 180% increase in unaided awareness for brands.

So, meaning you don't prompt them to specifically think about a specific brand in a category and then they are exposed to these inksyn media formats, you're seeing massive lifts in overall awareness.

We're also seeing things like on an entertainment program that advertised using rubberized NC media, we saw a 10% increase in the intent to watch that program by the viewers. Another campaign that we did for our big CPG brand, they saw a 19% increase in intent to purchase.

I saw somewhere, I think it was about 23% in brand favorability when considered against primary competitors. So these are real world impacts that brands are seeing by being in the content. And the reason why is because a, it was seen, it wasn't skipped.

B is it's on screen significantly longer. So at best you get a 32nd spot with our stuff. It's an average of about two and a half minutes. Some on screen is maybe ten minutes inside of a video.

But the third thing is you're also aligned with content that people love and that they're very passionate about.

And so a, when you see somebody, a brand, inside a creator content video, and you already love that creator and their content, there's a positive association because they're being supported by that brand. So you take a natural affinity towards that brand. But B, we can also track it in the comments.

Especially on YouTube and TikTok, you can see people talking about that brand and in some cases they're just like, hey, did you notice that ad? In some cases, they're literally saying, hey, I'm in the mood for such and such brand right now.

And there's a fair, unique opportunity to not only be able to track traditional brand metrics, but literally go and talk to the people, the consumers in the comments, literally jump in and engage with them and say, hey, if you're in the mood right now, here's a quick coupon, or here's your link, we'll send it to your house through Instagram.

Adam Boettiger: 00:33:40

Oh, especially on TikTok. Or Instagram videos, things like that. Sure. And you showed me some of those comments.

I think that people, of course, people don't want to be interrupted and nobody really likes ads, especially the interruptive ones.

That's why the television, the television ads, they had to crank up the volume to reach you while you're going to get bottled water or soda out of the refrigerator during the commercial break. So AI based brand injection into live or into a video library. Yeah, I think that sounds fascinating. And I feel like we're at the cusp of this.

I don't want to compare it to like in the fifties where we had the laundry detergent, people creating soap operas, but I think this is amazing stuff. It's fascinating.

Cory Treffiletti: 00:34:31

I think you're going to see the beginning of a full movement where non interruptive advertising and marketing and media.

Adam Boettiger: 00:34:38

Do you have any thoughts on the future of generative and regenerative AI? Where do you think it's heading? I've got my own thoughts, but you're here and we want to hear from you.

Cory Treffiletti: 00:34:49

I think regenerative is going to be an area of growth. Obviously, for us. If we've not coined it, then we've basically owned it and we're going to try to make that work.

I think overall, though, the generated AI and the AI tools, they're just supposed to make our lives a bit easier, to make our output more. More voluminous, more efficient, hopefully more effective.

I think it should enable us to do the analytics against what we're doing in a much faster turnaround time so we can continue to improve. I actually literally focus my time on using AI so I can be more creative.

And I think that's the core of how it can be done, is let's get rid of some of the basic fundamental stuff that needs to be done so I could be more creative. I want to be more creative.

Adam Boettiger: 00:35:32

Right? And yeah, that brought a question to my mind.

Going back to your insane media, how are people, it seems like that would be for more brand advertisers than direct response advertisers. How are people measuring the effects of engagement and Lyft and all of that stuff?

Cory Treffiletti: 00:35:53

I think in a basic thing is it's a matter of efficiency at exposure.

So the top of the funnel, it's a more effective CPM and time on screen in the middle of the funnel, to your point, it's more traditional brand metrics like awareness, intent, favorability, et cetera.

But there are lower funnel metrics that we also can employ too, where we are capable right now, through certain partnerships, to be able to look at lift in sales against the exposed audience and then scale that, because we can do a billion, for example, video views per month. So we're tracking sales lift against an exposed audience of maybe, say, a few million people from a brand.

But you can scale that to be able to reach almost the entire audience for a brand and then look at the overall sales lift. So we have those tools in place to look at bottom of funnel as.

Adam Boettiger: 00:36:38

Well as an example of your stuff, I saw a really cool advertisement, and it was for Tide, the tide, white, penniless. And you guys had created a little pen, but animated it.

You see the video going on with the influencer talking or whatnot, and then you see the tied pen on the table and all of a sudden it just pops. Pops, and then jumps up onto the white shirt of the influencer and erases a stain and then jumps back down on the table.

I thought that was pretty cool.

Cory Treffiletti: 00:37:05

Yeah, they're storytelling placements.

Adam Boettiger: 00:37:08

Yeah, there's no way to ignore that. And I looked at the comments and people were just loving it. It was just like, wow, this is not interruptive.

And also from an advertiser standpoint, I thought, because putting my CMO hat on for a minute, this stuff doesn't require four weeks of lead time to produce a video commercial. It doesn't require huge production cost. So I thought that was fascinating.

And in terms of just three things that you would have practical advice for people listening today who are in marketing or in business looking to integrate either insane media, but preferably generative AI, AI tools, things like that. What would you tell them if you could speak to them directly?

Cory Treffiletti: 00:37:56

I'll tell you exactly my answer, and my answer is one level higher than just on Genai, is that if you want to be successful these days, there's four things you got to know how to do. You have to be able to be strategic. And I think everybody at a certain level is strategic.

And if you're a not at that level and you're continuing to grow to it, you're learning, you're going to continue to be strategic, have to know how to manage people, right. That's a no brainer in the workforce. You have to be able to build relationships across an organization.

So if you're marketing, that means you're talking to sales and product and other groups. But then the fourth thing that everybody forgets is you got to know how to do stuff.

And that just basically means don't think your day is over because you managed a whole team to get stuff done. You act strategic and you sat with a bunch of people from different groups inside your company.

You got to make sure you spend at least a portion of every week doing stuff.

So get into the systems and mess around and learn how to use them and break them and try to fix them and do stuff because you cannot get caught on your back foot anymore and expect to be successful because you didn't know how to do something.

Adam Boettiger: 00:38:55

Yeah, I totally agree with you. I'm speaking with other peer chief marketing officers who have as much experience as you and I do.

Everybody's finding that there's startups and brands and companies that are looking for people who are not just strategic management leaders but are actually tactical leaders. They can do the stuff. They can walk the talk. And I think it's important.

Continuing education was never as important in the past as it is right now today because this is changing so fast and there's so much information out there that you just need to figure out which sources you want to go to for the information, how you want to consume it. But it's all about learning on an ongoing basis.

And like I tell my sons and everybody that listens to me that you really have to become a lifelong learner. You have to.

That's one of the reasons why I went back to graduate school at my age for the MBA is things changed back in the nineties, early two thousands. If I had gotten my MBA back then, much of the information probably would have been outdated anyway. Corey, thank you so much for joining us today.

I want to be respectful of your time, but we appreciate it. And how can people find out more information about Rembrandt? And how can they connect with you or subscribe to your newsletter?

Cory Treffiletti: 00:40:20

The easiest thing is go to rembrandt.com and you can sign up for information there. You can shoot me an email, Corey. Cory renbrand.com and you can also look@originaldigital.org and you'll be able to see some stuff there too.

Adam Boettiger: 00:40:34

Great. Thank you so much, Corey. We appreciate you taking the time to join us today. And this has been the powered by AI podcast. I'm Adam Bodegar, your host.

Thanks for listening. And don't forget if you found value in this to hit the subscribe button down below and join us for the next one.

You can also provide feedback through the links in the show description. Thank you.