In the third episode of the "MINT Eyes On" series, Carlo De Matteo, Co-Founder and Chief Operating Officer of MINT, explores the transformative potential of generative artificial intelligence and advertising automation
Category
🗞
Novità Trascrizione
00:00This is Mint Eyes On, a series produced by Engage for Mint.
00:07The series explores the perspective of Mint, a global company specialized in advertising
00:13resource management, on the future of the media and advertising landscape.
00:19In this third episode, we will explore the future application of JNAI and advertising
00:25automation with Carlo De Matteo, Co-Founder and Chief Operating Officer at Mint.
00:35Welcome to our new episode of Mint Eyes On.
00:38Carlo, thanks for being with us today.
00:40Oh, very much a pleasure and thanks for having me here.
00:44Carlo, let's start with the elephant in the room.
00:47The advertising industry has long been characterized by very complex processes involving multiple
00:54actors and technologies.
00:57Despite the clear need for efficiency, it's still difficult to find a unified platform
01:03that can automate all the processes.
01:05Why, in your opinion?
01:06The media industry has indeed a very complex process, but it's not sort of like a standardized
01:12process.
01:13So, you have a variety of processes, you have a variety of players in the industry.
01:18You have the advertiser, you have the media agency, you have a publisher, you have auditors,
01:24you have a...
01:25It's very multifaceted.
01:27And historically, the process was...
01:31It is still manual.
01:32It was manual.
01:33It is very still manual.
01:34What happened is that it emerged over time some point solutions to tackle some of the
01:44process steps of a gigantic and cumbersome process.
01:49And this is actually something very natural when you discover out in an industry how to
01:54fix the hurdles of people working with manual processes.
02:02And it gets to a certain point in which someone has an intuition, like, let's try to make
02:08all the players and all the processes within one single unified platform.
02:13And this is very much, I guess, a matter of timing.
02:16So, I think that the industry has evolved technologically so much that at the same time,
02:24there is the possibility to develop and to start introducing a technology that really
02:30offer the overall scope of the advertising process into one single solution.
02:36All right.
02:37Now, let's shift our focus on two hot topics of the advertising industry, automation and
02:45artificial intelligence.
02:47How these two technologies are impacting on the advertising processes?
02:53Well, very much, although we are at the very, very beginning of the long journey that we
03:01have ahead in this field, perhaps a little bit more advanced in the automation phase.
03:08Automation and AI are often put together, but in reality, they're not quite the same
03:13thing.
03:15So, at least you can have automation without AI, you most likely want to have automation
03:18if you have AI.
03:19Let me explain this a little bit better.
03:22So, in advertising, it's been quite a while that you have some sort of automation here
03:27and there.
03:29Automation use cases are very frequent, like you have automated report, for example, from
03:34many companies.
03:35So, that's not something that surprise anyone.
03:37But when it gets to start automating the process, that is where you don't have anyone offering
03:42it.
03:44How do I automate from my media mix model to my media plan?
03:48That doesn't exist.
03:50How do I automate my media planning to my buying plan?
03:53And so on and so forth.
03:54So, all those steps that can easily be automated through a software today still live in separate
04:00silos and managed by different people.
04:03So, automation in this field is very much, obviously, has to be software-led.
04:09Because software is the only one that can really automate the process.
04:12Now, if we move in the space of AI, then we need to be a little bit more, let's say, academical,
04:18if you wish.
04:19There is the, it's fine to say traditional AI, AI is not something very old, but the
04:25typical AI machine learning algorithm that you will find in advertising, which already
04:29exists for, you know, since a while, it's, they're mainly focused on optimization.
04:37So, you would want to have a software powered by machine learning, powered by AI, to offer
04:44you different scenarios where you can optimize your spend across channel or within the same
04:49channel.
04:50You know, there are a lot of powerful optimizer out there in the industry.
04:55Now, what's the last one in AI space, is the LLM, is the large language model, you know,
05:03we've always been hearing about charge BD, co-pilot, et cetera.
05:07And again, here we have two different time, you know, you have the typical natural language
05:12conversational tools.
05:15And in advertising, these type of tools can really help brands and agency to cut off a
05:23lot of work by just simply creating a prompt, asking something and discovering data and
05:31explore the realm of your data set in a nice and intuitive way with charts and pies and
05:38et cetera, whatever you like.
05:39Okay.
05:40So, that's the conversational one.
05:41The most advanced one, which is the one I like the most, is the concept of a multi-agent
05:49or synthetic agents.
05:51So, you can have agents that, for example, are deployed in a development team.
05:58So, someone can help you to do a front-end development or back-end development or QA.
06:04But the same agent can be applied to media.
06:08So, a team with a media planner and digital specialist and, you know, a trafficker can
06:14work alongside a group of agents that will help them to do the job easily.
06:20And will do the most difficult part, most long-time part in a second using the technology
06:28of LLM.
06:29And this is really the future.
06:30And that's where Mint is also looking at that space.
06:34So, clearly, there are a lot of potential and advantages in using these technologies.
06:42Still, it looks like companies are kind of struggling in adopting and implementing them.
06:50Why is that so?
06:52As you know, when some new technology hit the market, there is a huge hype.
06:55Everyone talks.
06:56Everyone is an expert.
06:58Everyone has already adopted since day one.
07:02It's normal.
07:03We expected that.
07:04In reality, not many of them really adopted on a daily basis.
07:07And adoption curve is, you know, sometimes it's quite steep for people who are not used
07:13to it.
07:13So, but this is just a matter of time.
07:15People will start using it, learning how to use it, and will become part of our everyday
07:20life job.
07:21So, having an AI that is applied to data that doesn't match the reality is very much a useless
07:28AI.
07:29You won't be able to satisfy your needs if you don't have a foundational architecture
07:33of data that is consistent with the reality.
07:37So, that is why companies start adopting tools like that.
07:42But in reality, they're not really making a lot of use of it because they need to go
07:46back to the report in Excel and make numbers right.
07:51So, fix the data first before starting with what you can put on top of your data.
07:57Now, what's the role of Mint in this evolving landscape?
08:02What are you doing for addressing these challenges?
08:05Well, first and foremost, our software is a collaborative software that stitch the process
08:12of the agency and the clients all together in one single solution.
08:16So, that's the first answer we give to the market.
08:19And that's what we've been doing, you know, since the beginning.
08:21Furthermore, what we are really focusing on is the concept of data harmonization and
08:29taxonomies.
08:30We discovered that this is really one of the biggest pain points of any brand.
08:36They struggle to make sure that the way in which they work with different data sources
08:43and they call things in a different way and they try to, you know, stitch the logic of
08:47the complex ecosystem.
08:49You know, most of the brands have multiple products, multiple brands, different geographies,
08:53and that's where they struggle.
08:54Taxonomies, not anything else.
08:57And that is a strong piece of work that Mint has been really investing a lot of money on.
09:04Making sure that you can automate, you can make sure that you can streamline, that you
09:08have a high governance on data harmonization and taxonomies.
09:12That's definitely something we have been leading, let's say, in the industry with.
09:18The last one is the Gen AI.
09:21I mean, Gen AI is not just a marketing word.
09:23It's very much the future.
09:26We, again, we live in an industry where processes are very much unstructured and there is a
09:35variety of manual work and repetitive tasks that should be reduced to a minimum.
09:42And this is where AI really comes in.
09:46The new investments we have are more in the space of Gen AI and LLM.
09:51We are going to launch very soon.
09:54Our co-pilot, our Mint co-pilot, where you can basically create your prompt, ask a question
10:02on your data set and have all the answers you are looking for.
10:07And in the future, we will be focusing more on creating those multi-agents that really
10:15are relevant for our industry that will help brands and agencies do their work in a better
10:21way.
10:22All right, Karol.
10:23Thanks a lot for these useful insights about the future of advertising.
10:30And it's been a pleasure to have you in this interview.
10:33My pleasure.