AI in Marketing: Real Results

Guy: [: 00:00:00

Relevant. Dot com. Today I'm interviewing Boesh, Ramad I of Latent View Analytics. And, uh, before we get started, let me tell you a little bit about him. Uh, Boesh is the head of Marketing Analytics Center of Excellence at Latent View Analytics, and he's gonna share valuable insights on blending data and AI to transform marketing strategies and enhance customer engagement and drive measurable.

, so good to have you today. [: 00:01:00

Boobesh: Hi guy. Uh, nice to meet you and good to have you.

Guy: Yeah. Fantastic. So, uh, so before we get started, what is your backstory on AI and marketing? How did you get involved in this, uh, fun, fun space? Yeah. Um,

Boobesh: I have been in the analytics space for over a couple of decades now.

Uh, in fact, when I started working in analytics, at that point in time, analytics was not even a key word. Not many people were storing the data and, and not using the data for their, uh, business driving business decisions. Uh, financial sectors were the initial ones to start using analytics a lot more in terms of driving business to identify.

ld because there is a lot of [: 00:02:00

Before it was only the broadcast channel or the events or the ad that you see while, while driving in a highway, right, where there's not a lot of information. But with the, with the whole advancement in, in the digital channel and everybody started using internet, uh, getting into Facebook, getting into Instagram, YouTube, TikTok, all of the social media have helped generate marketers to not only focus on the off channel marketing, but also the on channel marketing.

And with the online channel marketing, there is a lot more data that has started coming in. And with that data, the marketers have become a lot more smarter in terms of. What are the kind of customers that I need to focus on? What kind of information are they looking for? What kind of usage that they are doing in our, in our organization, in the products that we are using, and how do we brand that?

re is a ton of the data that [: 00:03:00

I. In terms of how to effectively use the data to, to drive their campaigns. And the space has been continuously evolving, both from the analytics perspective and from the marketers in terms of how they market and, and personalize the content for their audience. So it's been a, it's been a great journey.

Guy: Yeah, it sure has. Uh, it's definitely pretty crazy what's going on. So now with the, uh, the advent of, uh, of gen ai, how do you. C or maybe what are some practical use cases of how some of the Fortune 500 companies are using, uh, generative AI in their, in their marketing?

Boobesh: Sure. The marketing [: 00:04:00

To running market research campaigns, like running a survey or, uh, running a social media campaign. Uh, those things to understanding in terms of what the customers are looking for and which segment of the customers, what kind of messages needs to be passed on and, and in which channel it needs to be passed on to.

How do we measure, uh, some of this in a more real time and also kind of going past and doing personalization at scale, right. In all of this, there is a lot of AI that can play an important role. And all of this has been traditionally, the marketers have been using for the last five to 10 years. Right. And the data has been a key, uh, part of this decision making in, in each of this.

media in terms of what is a [: 00:05:00

And if you think about the market as a generally, it is a very unstructured data. Um, the, the focus group conversation is all voiceover, whereas the survey is all text format. The social media is all text format. And, uh, historically the companies which didn't have a bigger team to. To make sense of all of this data can now do it through gene Ed.

All they need to do is pass on all of this information into a model and they'll be able to ask questions and, and find out in terms of what is working or what is not working. So that is in terms of the brand perception and understanding. The second is in terms of the experimentation. Now, the one big advantage that I see the market as unlocking with the whole AI is.

especially the marketers to [: 00:06:00

Right. And the creatives can be created on the fly using some of the, which typically, uh, marketers will work with an agency team or a creative team within their organization, which will take anywhere from two to four weeks. The, if you think of it, the historical way, the marketers will go to the creative agency and say that, Hey.

Tomorrow I'm, I'm going to do a, a back to school campaign in the next quarter and in the back to school campaign, I wanted to come up with a, with a creative that is, which is more present, but also carries some of our past campaign performance, also carry some of the culture or the kind of product that the organization stands for.

. The marketers looks at it. [: 00:07:00

But now with the whole advantage of gene ai, all of this can be fast process and it can be done within a week. The marketers will be able to put together certain prompts to any of the, uh, LLMs that they have internally, and they will be able to kind of build out the first. Creative draft and then use that for announcing a lot more.

Right. So that is a, is a whole new unlock that I'm seeing for the marketers, uh, which in the coming days we will see most of the companies will start adapting it. And for this, you don't need to have a big budget, you don't need to have a big team to, to enable all of this. All of this can be done at at scale.

Yeah. And scale is the other thing that AI is going to unlock a lot more. Yeah.

reative you can be and, and, [: 00:08:00

Where you can prove its effectiveness. So, uh, what are some of the challenges that the organizations are facing today with AI and, um, maybe some of the, uh, you know, this term agentic AI systems that are out there. Uh, what can they do and, uh, you know, how are they moving from, you know, pilot into full production?

Sure.

Boobesh: So the biggest challenges that I see that I've noticed when we talk to some of the clients is, is, is the data, right? Because there is data algorithm and the business use case, the business use case, pretty much most of the marketers know in terms of what are the problems that they wanted to solve the algorithm before it was.

by a smaller company. Right [: 00:09:00

The third part of it, which is the foundation is the data, right? And the data becomes unique for every organization, and you cannot use the same data where another organization uses. You need to start capturing all of your customer's data, what your customer likes and don't like. What kind of product are they using, what kind of feature within that product do they like?

What kind of campaigns are they responding to and where you are getting on a higher. Typical metrics like open rate, click rate, impressions and all of that. So there is a DNA for every organization and that DNA comes from the product that you're offering and the customers who are using those product.

it. But when they go into in [: 00:10:00

The real challenge comes in and, uh, when we start having conversations with our client, everybody excited to get to a pilot mode, we get into a pilot mode, and then we discover saying that. Even though the pilot is a successful, for a particular champion who is coming in for a use cases, if you really want to scale it, we need a large variation of the data and that large variation of the data is not there in many organizations at this point in time.

In fact, that is where we also do a discovery access and tell them that not all data is useful, but what are the data that you need to capture and build your foundation models now.

Guy: You know, it's so interesting when you say that that variation in the data, you know, makes such a difference. And, and, uh, you know, and, and I, I think you're right.

So, um, how are, uh, larger [: 00:11:00

Boobesh: Yeah, yeah, sure. I think it is very sexy to say that every organization is working on ai, but if you really look at it in terms of what is, what number of organizations have moved. Their pilot project to production. The numbers is astonishing, right? Uh, in fact, the Gartner came up with a report saying that 83 percentage of the all csuite level wanted to work on AI and wanted their enterprise organiz team to, to focus on AI solutions.

ee from our own experience a [: 00:12:00

Out of that hundreds of conversation, one third of them saying that, Hey, this is great. I like what you're saying. Let's go and do a pilot. And the pilot is, takes anywhere from four to six weeks, we are able to do it. And from that pilot to production, we see again only one third of that moving into a production.

So there is a lot of filtering that happens and there are multiple reasons to that. Right. What we have learned from organization which are successfully able to move a pilot to a production is one, they do have a very robust data ecosystem. Or at least they are able to acknowledge saying that this is the data that I have today and this is the kind of data that I need to start capturing in the future.

of in. So it is not only a. [: 00:13:00

So the CMO is then able to tap onto the C-I-O-C-T-O of the organization and even the infrastructure team because we need to start bringing in the infrastructure team and the infosecurity team right there in the pilot mode to understand. What kind of components that we have been used and whether it has all been approved by the internal organization, especially when you're working with a Fortune 500 company where there is a, so much a protocol that they go with.

different functions to start [: 00:14:00

The third most important thing is we need to have a champion from the client, stakeholder who are willing to. Try and fail and also help in terms of improving. What I have noticed over time is we are very relaxing when it comes to human making a mistake. We say that, hey, that's, I mean, it's end of the day, it's a human and human will bound to make mistakes, but very, we are very critical when it comes to, yay making a mistake.

We don't want the missions to make a mistake. And we are expecting it to work like a calculator. Hey, if I say five plus five, it should give me an answer. Why give an another? But that will never be the case. Uh, when, when you are, when you're starting up something, there is obviously there is gonna be some level of hallucination, some level of fine tuning that we have to do.

at is, uh, that is the three [: 00:15:00

Guy: Yeah, absolutely. And, uh, you know, and, and, and it's kind of funny because, you know, we do expect AI to be perfect and, and it's so funny when you see it hallucinating and going off the rails and, um, but you know, it, it's, if I think about even, you know, how AI as. Being used in driving a car. If that AI driven car has an accident, then it's AI's fault.

lot more mistakes than, than [: 00:16:00

And, and I think it's, uh, pretty much the case. Maybe not entirely yet, but it's moving there where. Yeah, humans are still making more mistakes than, you know, AI generated, uh, results and stuff like that. So, uh, what do you see as, uh, maybe the top two, uh, best practices now in using Gen AI and some of the larger, uh, companies that you deal with?

I.

Boobesh: I think one is definitely there should be a buy-in from the CSO in terms of pushing their team to become an AI first approach. I think all of us as an individual need to start spending time and, and using AI in, in whatever the capacity that we wanted to use. I read somewhere it says that the next generation of workers who are going to come into the, the professional life are going to be AI first.

ady showing all the signs of [: 00:17:00

Another thing, they are not going to do it the the old fashioned, they are going to do it in a much more digital friendly format and AI friendly format, and the models are going to get smarter. And when these people kind of come into the workforce. It is absolutely needed for all the organizations to be AI first company, and if in fact, there are some media and entertainment companies which says that, Hey, we will not be in this AI at this point in time because of whatever is happening, right?

In terms of there is a fight with the writers and, and there is, there is also, there are organization which says that. Maybe we will not take this route, but I think in the end, I think every company is going to become an AI based company or start using AI to not only improve the efficiency and the productivity, but if they really wanted to give.

compared to some of the past [: 00:18:00

We have all know in terms of what is a waiting time that we have in our, in a call when we try to call a customers, right? And if you are able to get a response in like few hours, then that is the kind of a company that we all wanted to go and partner with. Saying that, hey, there is one company which is able to respond to me in five minutes.

Either is a human or it's a computer. Whereas another company, I need to stand in line to talk to a human for like two hours. Maybe I'll, I'll choose this, right? So the value to the customer is going to be the far more important and. If AI is enabling you to provide that value to the company, I think you should, you should start doing it, as you said.

We need to find out what are [: 00:19:00

Guy: Yeah. Yeah, definitely. So, uh, I'm sure you've got some, uh, personal favorites in terms of some of the tools, uh, that you're using. Which ones, uh, kind of, you know, really stand out, uh, for you as some of the things that you're doing in, in, in your business?

Boobesh: Yeah. The number one for me is the Notebook, lm, because, uh, I am a huge, uh, podcast listener.

And, uh, now I don't need to go through, uh, tons of research documents or, or a new publication article that comes out. I just take it out and put it into Notebook, lm and then it converts into a podcast for me. And that podcast is something that I'm able to hear when, when I'm doing. Whatever, walking or working out in the gym or driving a car, right?

that I have been, obviously [: 00:20:00

That is something that I have been using more regularly, but I've also been exploring something like napkin doi that kind of helps in terms of slide preparation, cursor, doi when I, when I really wanted to get into coding. Uh, some of this has been my personal favorites.

Guy: Okay. Yeah. Fantastic. Um, uh, that's, uh, you know, it's always helpful to, uh, hear what other folks are using and, and figure out how, you know, and try 'em out and see what's going on.

So, uh, do you see any differences in different industries as to how gen AI is being used for certain things in one industry versus another? I.

PG vertical has is when they [: 00:21:00

It takes anywhere, typically from two years to three years. Uh, the whole ideation in terms of what does the customers want. Two, in terms of what kind of ingredients can be put into a, into a makeup, uh, product, how the product needs to be packaged, and how do we test out this product and then going into the market in terms of how it needs to be positioned.

Right. All of this as a journey. It'll take like good two to three years. And, uh, even though they are able to find out in terms of what the customer needs and what kind of ingredients that they're looking for, for a bigger organization, if it takes three years, then the smaller influencer based. Smaller companies are able to do it in two, three months.

using the GNI to bring this [: 00:22:00

In the financial sectors. We have already seen a lot of use cases around the customer support. Wells Fargo has came up with Forge, uh, as a, as a support agent, where you can go and ask any questions about your, your account and, and the investments plans that they have. We have seen plan comparison as a, as a good use case if I wanted to go and look for.

What are the different investment plans that is available and how do I compare and contrast different between different plans or what are the different insurance plans that is available and how do I compare and contrast between different plans? Right. The plan comparison is a great use case that I've seen in the financial sectors, financial and insurance sectors.

ot more on the creative part [: 00:23:00

Cross channel attribution and using agents to go ahead and launch some of these campaigns has been a great use case. So it's been pretty interesting in terms of how across different interesting industries they are focusing on a particular function and, and going deeper on it.

Guy: Hmm. Yeah. Very interesting.

And, um, yeah, and it, it is funny how, you know, for a larger organization, you know, a CPG, a big CPG taking two to three years and, and the smaller companies can turn something around in two or three months. And, and there's definitely an opportunity for the, uh, these larger. Consumer companies to, uh, reduce that two or three years down to, maybe not two or three months, but certainly something significantly cheaper.

nd, uh, you know, so this is [: 00:24:00

Boobesh: A hundred percent. Yeah. Yeah. Yeah. And, and if you are not doing it, there is another competitor who's sitting by the side of yours and who's doing it and taking all the customers.

Guy: Yeah. Can you imagine, you know, you're sitting there and, uh, you know, you're, you've got your development plan, your product cycle.

It's two or three years away at your competition, is launching something today. And then, uh, six months later they, they've launched the next one and the next one, and you're still waiting on, you know, three years to go by before you're done. I, uh, yeah. Uh, big, big, uh, opportunity for, for them to, uh, significantly improve and, and, and accelerate that development lifecycle.

with ai, well, generative AI [: 00:25:00

Boobesh: Um, it depends upon what kind of marketer that you want it to be. There are, if you wanted to be a channel marketing, then you need to focus on a particular channel and find out in terms of what kind of agents those channels are building. For example, if you're focusing on Facebook or if you're focusing on YouTube, YouTube is going to come up with their own agents in terms of helping the marketers to do it.

Having a hands-on experience on some of that will definitely help. If you wanted to be become more of a campaign, uh, marketer where you pick a particular campaign, like holiday campaign or a theme-based campaign, and you wanted to understand how this campaign has been performing across different channels, I.

ind out which channel I need [: 00:26:00

If you wanted to become a campaign marketer, if you wanted to become more of a, a measurements and insights. For the marketing team, then there is a lot of brand insights and, and the measurement metrics that you need to be aware of I think have a sense, having a sense of, in terms of where all of this is moving towards and there is always new channels that that keeps popping up.

We see that, especially in the B2B world. LinkedIn is becoming a lot more powerful channel, uh, compared to some of the other things. So we, we need to understand, and even within the channels also, and every channel wants to keep their audience within their platform and not really have a click through and, and send it.

itely help you. And focusing [: 00:27:00

Guy: You know, it's interesting that you say that because, um, you know, one thing I've found with, uh, when I interview, uh, you know, entry-level marketers, uh, they really don't have a, a, a good understanding of what are all of the different areas that that could be, you know, be interested in or wanna work in. And to some extent, you know, all they want to do is just get a job.

But you know, one of the things though, I think where AI can really help them to hone exactly what they can do and what they want to focus in on when they get that fo that first job. So, you know, you mentioned a couple things. You know, whether it was brand insights or measurement or creative or whatever it happens to be is, you know, I, I think.

e AI to now really help them [: 00:28:00

Boobesh: Yeah, no, totally agree a hundred percent. I think that will also help you to keep learning on that particular space over time, which eventually you'll become a master of that space.

Guy: Yeah. Yeah, absolutely. Well, Boesh, uh, thank you so much. I wish we could go on, but, um, unfortunately, you know, doing a podcast for two or three hours, uh, some people might listen to it, but, uh, I think we're gonna have to close it off here.

Thank you again so much. It's really been interesting and you certainly have some great, uh, perspectives on, uh, on the gen AI space and marketing. So where can folks reach you and learn more about your company and go from there?

Boobesh: [: 00:29:00

I personally enjoyed, uh, this last 20, 30 minutes. I'm very easy to find out. You can go to LinkedIn and look for esad. I'll be the only person who will show up there. And, uh, you can reach out to my company through latent view.com. L-A-T-E-N-T-B-I-E-W.

Guy: Fantastic. So layton view.com. L-A-T-E-N-T-V-I-E w.com.

Boesh, thank you so much. I really appreciate it. And, uh, look forward to possibly, uh, chatting again. And for, uh, the audience. Please stay tuned for many other videos in this series of the backstory on marketing and ai. And please go to marketing. machine.pro relevant.com and download some, uh, new information on my upcoming book, the uh, AI Marketing Machine, Boesh.

Thank you so much.

Boobesh: Thank you so much, guy. All the best for your new book. I'm looking forward to it.

y: No, thank you. Appreciate [: 00:30:00