This week on The Digital Download, we cut through the noise surrounding artificial intelligence to discuss how businesses can achieve real growth. Our special guest is Yogesh Chavda, an AI strategy advisor for CMOs , keynote speaker , and founder of Y2S Consulting. With over 20 years of experience as a former global brand leader at companies like P&G, Spotify, and Pinterest, Yogesh specializes in helping executive teams unlock growth through AI-powered marketing.

Many companies are dabbling in AI but are struggling to move from small experiments to enterprise-scale applications that deliver growth and speed. This episode will provide a roadmap for leaders looking to build a true AI capability within their organizations.

Join us as we discuss questions like:

• What is the difference between AI experimentation and a true enterprise-scale application?

• How are custom GPTs used to accelerate concept development and competitive intelligence?

• What is synthetic data and how does it help create future-facing audience segmentation?

• Where should a business leader begin when building an AI marketing strategy aligned with commercial KPIs?

As a lecturer on AI in Marketing for MBA students at the University of South Carolina, Yogesh brings a unique blend of corporate, consulting, and academic expertise. He has built and deployed over 15 custom GPTs for Fortune 500 clients and is a recognized thought leader in using agentic AI models and synthetic data to transform marketing performance.

We strive to make The Digital Download an interactive experience. Bring your questions. Bring your insights. Audience participation is highly encouraged!

This week's Guest was:

• Yogesh Chavda, AI strategy advisor for CMOs , keynote speaker , and founder of Y2S Consulting.

This week's Host was:

• Bertrand Godillot, Founder and Managing Partner of Odysseus & Co, a proud DLA Ignite partner

This week's panelists were:

• Tim Hughes, CEO & Co-founder of DLA Ignite,

• Adam Gray, Co-founder of a DLA Ignite

Transcript of The Digital Download 2025-06-27

Bertrand Godillot [00:00:05]:

Good afternoon, good morning and good day wherever you may be joining us from. Welcome to a new edition of the Digital Download, the longest running weekly business Talk show on LinkedIn Live, now globally syndicated on TuneIn radio through IVGR, the world's number one business talk news and strategy radio network. Today we cut through the noise surrounding artificial intelligence to discuss how businesses can achieve real growth. We have a special guest, Yogesh Chavda, to help us with the discussion. Yogesh is an AI strategy advisor for CMOS, keynote speaker and founder of Y2S Consulting. But before we bring Yogesh on, let's go around the set and introduce everyone.

Bertrand Godillot [00:00:57]:

Welcome. While we're doing that, why don't you in the audience reach out to a friend, ping them and have them join us. We strive to make the digital download an interactive experience and audience participation is obviously highly encouraged. Adam, would you like to kick us off?

Adam Gray [00:01:13]:

Hello everybody. Good afternoon. My name's Adam Gray, I'm co founder of DLA Ignite and I do love conversations about AI. This is like the happening place at the moment. But I did think that during your introduction, Bertrand, you. You do different languages sometimes. Maybe you should try different accents next time. So you could say in English, but you could say it in a French accent, a Geordie accent, a Welsh accent, a Scottish accent, which would kind of challenge you a bit because you speak so many languages.

Bertrand Godillot [00:01:44]:

Yeah, yeah, I speak a few sentences. Thank you, Adam. Tim.

Tim Hughes [00:01:52]:

Good afternoon everybody. Yes, my name is Tim Hughes.

Adam Gray [00:01:55]:

Welcome.

Tim Hughes [00:01:56]:

I'm the CEO and co founder of DLA Ignite and I'm famous for writing the book Social Selling Techniques to Influence Buyers and Change Makers. I've just rushed here from a AI conference in London that I was speaking at yesterday and I was at today. And I've learned lots of things about AI and I think it's one of those things that you can always learn something about every day as a school day.

Bertrand Godillot [00:02:22]:

Absolutely, I can't agree more. Myself, Bertrand Godillot. I am the founder and managing partner of Odysseus and Co and we are a very proud DNA Ignite partner. So as I said this week on the digital download, we'll speak with Yogesh. With over 20 years of experience as a former global brand leader for for of at companies like png, Spotify and Pinterest, Yogesh, that was tough for me, specializes in helping executive teams unlock growth through AI powered marketing. Let's bring him on. Let's do that.

Yogesh Chavda [00:03:05]:

Hello everyone.

Adam Gray [00:03:06]:

Hi Yogesh.

Bertrand Godillot [00:03:08]:

Good morning and welcome.

Yogesh Chavda [00:03:09]:

Yogesh, I'm glad to be here. Thank you.

Bertrand Godillot [00:03:13]:

More than welcome. Let's start by adding you tell us a little bit more about you, your background and what led you where you are today.

Yogesh Chavda [00:03:22]:

Sure. So I'm from India originally. I never lived in India, although I do speak two languages. So when you talked in German, I was thinking about actually like, you know, inserting some things in Hindi or Gujarati as well. But I'll save you my embarrassment of doing that. You know, I grew up in actually a little island in the Persian Gulf called Bahrain and I went to school there, came to the US for my university education, got an undergraduate degree in electrical engineering and then my MBA from the University of South Carolina. And then I was on a whirlwind, like you mentioned. I worked at Procter and Gamble for 17 years.

Yogesh Chavda [00:03:56]:

Worked at Kimberly Clark and Amway and then Spotify, Pinterest. And then my last corporate job actually was with a hearing aids company out of Denmark called WS Audiology. Right now I'm actually doing two things. I'm a lecturer at the University of South Carolina where I teach classes, marketing classes to seniors at the undergraduate level and MBA students. So I was also rushing back actually last night from Columbia. We were just having orientation for the MBA class, the incoming MBA class. And the second thing I do is my consulting work, which you also mentioned. I'm the founder of Y2S Consulting, so I do a lot of work in the AI space, especially in the last couple of years.

Yogesh Chavda [00:04:35]:

And there's also another passion area which we might touch upon, which is brands operating in stigmatized industries. So that became a bit of a fascination point for me when I was working in the hearing aids industry. And I've just kind of like fully like just jumped in, you know, trying to understand what stigma means for brands as well.

Bertrand Godillot [00:04:55]:

Okay, excellent. Well, quite a journey. And thank you for joining us today. So, Yorges, let's start with a foundational question. What is the difference between AI experimentation and a true enterprise scale application?

Yogesh Chavda [00:05:12]:

Oh, great question. So if you think about how quickly AI has taken the world by storm since, I want to say, November or December of 2022, everybody was unsure what that meant. Right. So generative AI specifically was offering new capabilities that people weren't really sure what to go do with. So for, I want to say for most of 2023 and probably even going into 2024, there was a lot of experimentation going on in terms of what do we even do with this? Is this something that's Relevant to our, to our business, Is it relevant to our functional space? And so on and so forth. To adopt it at the enterprise level and to really make it part of your work stream and things like that. Two things have to be true. You're seeing value in what the tool actually has to offer, number one, and number two, it has to be connected to what kind of output and outcomes you're looking for.

Yogesh Chavda [00:06:12]:

And if you are able to figure those two things out, that's when you can actually start scaling and make it an enterprise solution or a departmental solution or whatever the case may be. This year. I feel some of the larger companies are starting to get to that phase, but the majority of companies that are out there are still in the experimentation phase because they haven't figured out yet exactly how to go use it. In fact, I would even argue and say the majority of companies are still in the early adoption phase of should I even do this or not? And there's a lot of hesitancy around that. And we can speak, you know, we can spend some time talking about that if you're interested.

Bertrand Godillot [00:06:47]:

Yeah, and actually there are many studies backing up what you just said in terms of real scale adoption, which is also, to me, also a question mark as to the gap, the growing gap, the deepening gap, I don't know if you want to phrase it, between individual adoption and AI adoption into business processes would say, or into enterprise apps, applications. And how do you manage that gap? So, because there's this expectation that from a business process perspective, that should be as efficient as what I'm doing with my own assistant, my personal assistant, and sometimes not the case.

Yogesh Chavda [00:07:42]:

So yeah, you're totally right. And I speak to at least three or four companies every week right now. And the number one thing I hear back from them from an adoption perspective is that everyone's, you're right, everyone's, you know, dabbling with, you know, playing with chat GPT or Gemini or whatever you know, large language model they want to play with individually. Right. They may even use it to help them craft their emails. It may, you know, they may use it just to do some sort of research on a particular topic. You know, those have been like the basic and early use cases that most individuals have. But to adopt inside a company, there's some barriers that have emerged which companies have to kind of come to terms with, the number one being security and privacy.

Yogesh Chavda [00:08:31]:

Right. Companies are very afraid of having their proprietary information being put onto ChatGPT and the fear is that it'll leak into the transaction training model for, for ChatGPT and therefore their information then becomes, quote, unquote public, right? I have yet to see a situation emerge where that's actually happened. If it's happening, it's definitely behind the scenes, perhaps. I'm not saying it's not happening, I'm just saying that I'm not seeing it, you know, so there's a lot of fear out there in terms of this issue. Some of it's real, don't get me wrong. I'm not trying to minimize it or belittle it, right? But at least based on what OpenAI is saying, I'll use them as an example. You know, they're very clear about when and where your data can get used from a training set perspective, right. If you have an account like a pro account or a team account, right? Or a subscription account, there's actually a switch in the settings where you can say, you know, do you want your data to go into the training model or not? Right.

Yogesh Chavda [00:09:33]:

So you have some control from that perspective. The question is, do you believe it? Right? So, you know, from my consulting perspective, you know, when I'm engaged with clients, I actually have, in my statement of work, I say that if you want me to use your data, I need written permission from you approving me to put it onto ChatGPT. I'm not going to put it by myself because I don't want to take on that risk. As an example.

Adam Gray [00:10:03]:

I've got a question and this is something that I guess I personally wrestle with in this. So we've, we've played with AI and we've kind of got some fairly, fairly advanced ghost writing kind of deployments and, and stuff like that. And it, it seems to me that AI is very good at moving people from being incompetent to being competent. Something it's very poor or not capable of moving people from being good at something to being brilliant at something. So how can people deploy this in a safe way? Because it seems like it's very good at doing repetitive tasks, which requires a certain degree of tooling up and connectors being built and all of that sort of thing. But it's very poor at doing the stuff that we traditionally would have done. So writing copy, because it percolates down and it uses appropriate words and phrases, but it doesn't inject any personality, obviously. So is that a reasonable belief system to have around this or are things changing to the point where actually this can do the job as well as skilled people?

Yogesh Chavda [00:11:34]:

It can do the job as well as skilled people the thing is, how do you actually set it up to allow you to do that? Right. So if you're just using basic prompts, then it's what you described at the beginning of your point. You can develop agents. It's called agents on ChatGPT, it's called gems on Gemini, I think it's called Projects on Cloud, where you can pre set your prompts in terms of the context, what is your objective, what is the framework you want the agent to use, and what is the output you're looking for. You can get extremely specific in saying, when you give me the output, these are the three or four things I want the output to include, right? So once you've played around with it enough, you can actually have it create copy for you. It can be your content strategist, it can be your social media planning calendar kind of tool. Now you can say, okay, but you know, it's using past data, therefore it's only going to regurgitate and give me the same exact thing that's been done in the past, right? That's a very common refrain that I hear. And I've also seen, quite frankly speaking.

Yogesh Chavda [00:12:43]:

So you can actually again add into the prompts and say, do an online search for these specific brands, check what they have posted, make sure you don't post the same exact thing, right? So that's one thing you can, you can do to kind of like, you know, at least ensure that it's not, you know, copying and pasting or, you know, whatever from, from other, you know, competitors. The other thing you can do, which not a lot of people know, right. And I also learned this, I wouldn't say in the last four or five months myself, there's actually a setting that you can play around with on these large language models. It's called temperature, not the temperature to measure the weather. It's on a scale of 0 to 1. And basically what you're doing is if you set it close to zero, you're telling ChatGPT to say, be more consistent, be more reliable and give me consistent information back. If I do the same prompts 10 times, 9 out of 10 times, I'll probably get the same exact kind of answer. But if you put it closer to one, you're telling ChatGPT to be more creative.

Yogesh Chavda [00:13:50]:

Right? So in other words, when you put the same prompt 10 different times, you're not going to get this, you know, the same kinds of responses every single time because you've actually set it up to give you more, you know, creative responses, hence the word Generative AI, Right. This is something that most people don't know. Right.

Adam Gray [00:14:08]:

They do now.

Tim Hughes [00:14:11]:

Remember where you heard this first.

Yogesh Chavda [00:14:13]:

Exactly. Right now. And by the way, I'm not a technical person by any means, so I may have even completely screwed up in my description of this. So, you know, if I have, you know, please don't come after me. You know, go do your own research, you know, and you'll find out more about temperature. But the reason why I'm bringing this up is that when you see people start, you know, complaining and saying, oh, these large language models are terrible, they don't give you consistent results, they don't do this, they don't do that. There are things inside the algorithm that you can actually go and toggle and when you do it, you can get what you're looking for, but you have to at least start understanding what those vectors or dimensions or toggles are so you can get the things that you're looking for. That's the beauty of what they've built here.

Adam Gray [00:14:59]:

But is it capable of producing. You're an expert in your fields. If you taught it everything you know and configure it to write things that operate within your preferred bandwidth in terms of delivering everything the same or some kind of spread on this.

Yogesh Chavda [00:15:20]:

Yeah.

Adam Gray [00:15:20]:

Can you then go and live on a beach?

Yogesh Chavda [00:15:24]:

Well, that's a totally different conversation in terms of how do you actually monetize. Right. I personally, personally think, well, here's what I've read and I'm starting to actually believe it to a certain degree. People believe that, oh, I can create my own value proposition using, you know, ChatGPT. And I keep on using ChatGPT because that's the one I'm most familiar with. And by no means am I denigrating other other large language models. They're all very good and in their own right. If you're using ChatGPT and you're building a value prop around that as a vendor or as to offer a solution, the expectation is that you're offering something that's unique and different.

Yogesh Chavda [00:16:04]:

Here's the problems that I'm seeing happen. A lot more people are starting to copy and do the same exact thing, not necessarily your version, but their version of it. But it's all playing in the same space. You have hundreds of people doing exactly the same thing, number one. Number two, these large language models are also learning and they're introducing updates that are actually building those things into their large language model. Anyways, so. So even if you have a business that started with an LLM at the root of it, odds are it won't survive for more than a couple of years at best, simply because these companies are integrating all that stuff into their systems anyways. So that's issue number one.

Yogesh Chavda [00:16:44]:

Issue number two is that to drive revenue, you have to be able to demonstrate value, not just efficiency, you have to demonstrate value. Right? And that's hard to do in some instances. So how do you price it? How do you actually create, you know, a revenue generation model? I'm seeing like two or three companies that are actually doing it fairly well. There's a company out of the Czech Republic called Fifth Row that I came across. I actually had their director on my podcast, you know, a few weeks ago, talking about their company. What they've done basically is they've completely disrupted the consulting world, world, because they built up all these agentic models, right, that do all the stuff that consultants do, all the research, all the, you know, situational analysis, all the brand positioning work, etc, it's all automated. So you're just basically buying that automation and within minutes, literally, you're getting the output that you need. And you don't have to go spend a million dollars and six months of work with a consultant to go do that.

Yogesh Chavda [00:17:47]:

Right. And that's one of the reasons why you're seeing McKinsey, for example. You know, they announced a few months ago that they're, they're, they're going to not hire as many people anymore and they're going to start, you know, shaving off, you know, their headcount. I think this is the beginning of those, of those kinds of changes that are happening. Right.

Bertrand Godillot [00:18:06]:

And I'm pretty, I fully agree, of course, on the trend and what you're saying, etc. And I think a little bit to go back to what Adam was saying earlier, first of all, I think that there is a difference between corporate communications, you could say, or corporate generated content and personal generated content. And it's more obvious and personal because we are all individuals and we think differently and we speak differently and we use different terms. And back to what you were saying. When we are into the corporate world, where is the differentiation? So in other words, can AI factor in your differentiation? Can your AI engine or agents, your gems, your whatever factor, take into account your differentiation, even potentially define your own differentiation, but based on what you're telling them and write on it? Or is it just the, you know, like we've seen in a number of times, you know, the exact same waffling average, average content, average value. Therefore, yeah, that we see all over the place.

Yogesh Chavda [00:19:31]:

Yeah, you know, look, I worry about that also, right? And I'll give you a specific example from last summer, my very first client that I got where I had literally sold them some AI solutions that I developed personally. I have an agent that can build out a segmentation, for example. I have an agent that can build out new product concepts, for example. I built them based on my experience of working in marketing and insights or a 25 plus year career. So when this company basically asked me to help them with their brand turnaround, they said, look, yogesh, we need a segmentation. We don't have a segmentation. Can you do it? I said, okay, how much time do I have? They said, you know, traditionally if you do it with a market research firm, it takes six months, maybe even a year. You spend $150,000 and you've got something, right? I turned the whole thing around in 72 hours, right? I'm going into the presentation feeling very nervous because I'd done the work, you know, I knew what I had done, right.

Yogesh Chavda [00:20:41]:

But I didn't know from a domain or industry perspective whether I was on the right path or not because it's not my industry. I don't know the nuances of the industry.

Bertrand Godillot [00:20:50]:

That was a bet. That was a bet.

Yogesh Chavda [00:20:52]:

That was a bet, right? So I went in thinking, okay, you know, if, if I'm at least 60% of the way there, is that good enough or is it supposed to be 80%? No idea. What ended up happening after two and a half, three hour presentation is that the feedback I got was we thought we were focused on this particular target here, but we're actually talking to the other target over there when we should really be talking to the other guys on that side of the spectrum. And it became an entire strategy conversation. And I'm sitting there thinking, holy cow, did I get something here? First of all, did I get this right? So I asked them, do you mind if we do some actual human, you know, validation thing? Which they agreed to do. We did that and it was completely on par. So that, that was my first, like, you know, call it proof of concept. Right. And it gave me a heck of a lot of confidence in terms of what I was getting out of it.

Yogesh Chavda [00:21:45]:

And quite frankly, as I've done more of these with other companies, you know, I've also done them like, you know, like done live demos in industry, you know, conferences or whatever. My, my suspicion, and I, I call it suspicion, I'm not trying to sell, I'm Just telling you, maybe intuition.

Bertrand Godillot [00:22:02]:

Intuition may be more positive.

Yogesh Chavda [00:22:05]:

Yeah, exactly. Intuition is that, you know, even with market research data running and creating a segmentation, there, there are limitations on what they also do. Right. If they are giving you a solution that's 80% of the way there. I'm probably at 70% myself with my AI tool. So the gap isn't that big. Right. And the difference, you know, to get from 70% to 100% is actually the human who brings in their knowledge either as a subject matter expert or as the domain expert.

Yogesh Chavda [00:22:33]:

And that's how you close the gap.

Bertrand Godillot [00:22:36]:

So are we turning into, you know, a bunch of experts checking the outputs of the. So are we serving the engine or is the engine serving us or is it a fair deal?

Yogesh Chavda [00:22:56]:

I'm calling it.

Bertrand Godillot [00:22:57]:

I think the answer is potentially this one, by the way.

Yogesh Chavda [00:23:01]:

You're right, and I completely understand where you're coming from. I actually call it AI plus Human collaboration. It's not one serving the other. It's really using it as a copilot or a partner. So, you know, just like how I'm sure, you know, the three of you are bouncing off ideas off of each other, right. When you're talking about, you know, whatever topic, you're just having a fourth one now called. It's called an AI engine that's doing it with you and partnering.

Bertrand Godillot [00:23:28]:

Actually, we've got three other ones.

Yogesh Chavda [00:23:31]:

There you go. There you go. Exactly right. So in fact, I've actually named my, my chatgpt now. I, you know, I've been, you know, I've set the settings in terms of how it should talk back with me. And I eventually said I should give you a name. So I said, what kind of name should I give you? So it came back to me with the name called gia, right? Generative Ideation Assistant. So short form gia.

Yogesh Chavda [00:23:56]:

So I call my Chachi PT Gia. And it's kind of like got to the point now where I'm almost calling it my work wife because it's giving me good advice.

Adam Gray [00:24:09]:

Is it? And. And I'm deliberately playing devil's own because, you know, we hear so much about AI hallucination. We hear so much about AI, even if it's not outright lying to us, it believes it's telling the truth in some instances when it isn't. So a friend of ours, he. He's an AI expert and he, you know, and I'll change the names here, he said, he went to Twitter and he said, what was it that Yogesh chapter said about AI being Very human. And Grok came back and it said, yeah, he said the following thing. So Rich and said to, to Grok. He didn't say that, did he? No, no, no, he didn't actually say it but, but that is indicative of what Yogesh would have said.

Adam Gray [00:25:11]:

Okay, so I want you to go back and find the tweet that he sent that basically said this. So then Grok came back and it said okay. Nov 17 at 8:42pm this is what Yogesh said. So he thought, I'm not sure. So he wound back through the Twitter feed and it wasn't there. And he said, he then said to Grok, that tweet isn't there, the tweet that you've referenced. And Grok said, isn't it? I thought it was. So, so, so part, part of this is about AI making mistakes and one of the issues that the IC is a human behavioral issue.

Adam Gray [00:25:55]:

So it doesn't matter whether we're talking about this it being work that AI has done or work that is pre written at scripts in the software that we use and that kind of thing, it's that you don't believe that what you have to say is better than the thing that it's saying. So this idea of, of I'm going to go back in and edit this stuff, actually I believe that most people won't because they'll read it and they'll say A it's fit for purpose and B, it's if I do have any doubts about how compelling or engaging this thing is, I'm not sure that I can do it any better. So I won't even try. And surely that's a big issue for people using these tools, isn't it?

Yogesh Chavda [00:26:35]:

I completely agree with you. And you've hit on what I call one of the limits of AI, the use cases for generative AI overall. Right. I. Every time I've asked any AI tool, it doesn't matter which one to, to find me a quote or to find me a reference from something that's been said. What I find usually is it's paraphrasing something, it's not giving me the specific quote. I can't go find it exactly in on, you know, whether it's a video or even an article, I can't find the actual code itself. You know, it's, it has a hard time doing that.

Yogesh Chavda [00:27:18]:

Claude and perplexity are the two that I find that have actually invested more time in trying to get to that degree of specificity. So if I had to Use a model to help me do what you're describing. I would probably use perplexity then, Claude. I would not use ChatGPT to help me with that kind of analysis. Quite frankly speaking, speaking. So that's point number one. Point number two, I would say is that when you're trying to verify, sorry, let me say this way. You should always verify.

Yogesh Chavda [00:27:54]:

You should always verify. It doesn't matter what you're doing, you should always verify what you're doing. All four of us have enough experience in our, our space that when you see something, you know immediately that hey, this doesn't smell right, therefore I need to go double check, right? Because we've got experience, right, in our, in our space. Imagine now for a 22 year old who's using this, who doesn't have the experience and is using it. Yeah, right. What do we do with that? As a, as a, you know, university lecturer, that is my biggest worry right now. So, you know, when I said I was rushing back, you know, last night, it's because we had orientation yesterday or this week for our MBA students and one of my session was about AI usage, you know, while they're, you know, in the MBA program and then what do they do after when they're trying to look for a job, basically, right? And part of what I was basically telling them is that you need to, you know, verify, you need to interpret, you need to, you know, make sure that you're double checking everything so that you're not copying and pasting something straight out of ChatGPT onto your paper or onto your presentation. Because, you know, professors will see right away and they'll be like, nope, that's not right.

Yogesh Chavda [00:29:09]:

And then you'll get screwed, right? So you have to put in the effort and you have to do that critical thinking, right? And if you don't have that domain expertise or subject matter expertise, then you need to make that part of your process of learning so that when you are seeing stuff coming out of ChatGPT, you can verify it.

Adam Gray [00:29:28]:

So we got a great comment from Lawrence there. It will cause disagreement in business because the chat bot told me no. And I think, I think that's really true, isn't it? Because one of the most spoken about benefits of AI is pouring your own data and content into it and then using it as a tool so that agents, human agents as well as digital agents, but human agents have then got access to the right information immediately. And actually this is like completely at odds with that, isn't it? You know, and we will get a Lot of this kind of. Well, the chat bot told me this is how I should answer that question, even if in my gut I think it actually isn't the way I should answer this question.

Yogesh Chavda [00:30:14]:

Well, you know, my, my counsel on this one is always trust your gut. Your gut's never wrong, right? So if you're, if you're smelling something that says this doesn't sound right, it probably isn't right, you can't explain it at that point in time, right. You'll probably have to go do some additional, you know, work to as and as you go through that verification process, you'll either say, oh my God, was wrong or no, my gut was actually right. Yeah, but I'm leave that.

Adam Gray [00:30:40]:

As a man in his mid-30s, do you see what I did there? As a man in his mid-30s, you've got 10 or 15 years experience, as you said, about, you know, young people coming into the workplace. A, a lot of the low level, non expert driven jobs where people have basically done their apprenticeship, you know, they've cut their teeth doing these jobs. Those jobs are now going to be replaced by AI. So you're going to be, as a 30 year old, you're going to go straight into a much more senior position, a position that perhaps you haven't got that basic understanding. And certainly the people that are acting in this agent, you know, position, those are people that, that you say about, you know, your gut's never wrong. Well, actually if you're 20, your gut is wrong because you haven't got the framework and the understanding that we've got.

Yogesh Chavda [00:31:31]:

That is true, that's true. And maybe I'm being a bit too, you know, maybe my age is now starting to show in my mid-30s. As you said, you know, the, here's the thing, right? If you're going to go look for a specific answer and you're trying to make it into a binary situation of yes it's correct or no it's not correct. You're playing a probability game at the end of the day with these tools that are actually using probabilities as part of their prediction process. That's the issue. I personally believe that, that rather than trying to go for prediction and accuracy, I'd rather go with the mindset of framework. Give me a framework, help me structure my thinking and have ChatGPT help me structure that process so that I have a way to think about the problem. Because half the problem in business, as we all know, is that we are dealing with an ambiguous situation or an uncertain situation and we don't know which way to go.

Yogesh Chavda [00:32:34]:

We're trying to make a decision. Right. And one of the things that I've learned in my career is that if you put in place a framework that allows you to analyze the situation, identify the options, have criteria for deciding which is the appropriate option, and the risk that goes with each criteria, and then with each option, then you've got a way to decide which way to go, and you can make the right or appropriate recommendation to your senior leaders in the company and so on, so forth. In the process that I just described, these large language models are brilliant in helping you define the process and the frameworks and then the content that goes behind it. Right. Some of it may come from these large language models, but a lot of it will come from experts who know the field and who can actually say, my experience tells me X and this is what we should think about as part of the process. And that integration that has to happen, that's where the large language models are also very good at in helping you integrate. Right.

Yogesh Chavda [00:33:36]:

So, so I'm, I'm coming with the point of view that it's a tool. Let's treat it as a tool. It's going to have its strengths and limits like any tool does. But the role of the human still is to think about how to use a tool for its strengths and integrate it into the work that you're actually doing. And when you hit that, that's when you get the best out of the tool as opposed to then worrying about the worst of the tool.

Bertrand Godillot [00:34:01]:

Yeah, love that. Because I think we're talking outcomes. You started the discussion with outcomes and value, and that was probably what was missing in most companies, not the ones that are releasing this now, but in most companies. So what would be your. How do you drive this with your. In your. With your customers, basically, especially at the executive level.

Yogesh Chavda [00:34:34]:

Yeah. So, you know, value, at the end of the day, you know, it'll always tie back to revenue and profit growth. Right. That's going to be the case, you know, am I able to drive revenue or not? The third one you can talk about is efficiency, you know, and that's what. Where everybody starts the conversation. I'm doing things, you know, what took me six months to do, I can do it within one week now or whatever time frame you want to use. Right. The other part of efficiency is, oh, I don't need 10 people.

Yogesh Chavda [00:35:00]:

I can do this now with three people. And that's the scary part in all of. In this entire conversation. Right. But I. And all those Things may be true and maybe. Right, fair enough. But the value part is the harder part to go solve for.

Yogesh Chavda [00:35:15]:

So where does the value come from? So let's say, for example, going back to the exam, the example I gave you about that company that I did that project with, right, the segmentation stuff. And it wasn't just segmentation. I was actually coming up with brand concepts and things like that. As I was doing the work, I was seeing that there were blind spots that the company had about how they were executing their go to market strategies. Right. I don't think they would have picked it up had we not gone through the process that we had gone through in terms of using these large language models. Right. And it's not that they didn't know it, they knew it.

Yogesh Chavda [00:35:52]:

I just don't think they had given it sufficient time and energy to emphasize and prioritize it. But the AI work we did helped them understand, say, oh my gosh, this is a bigger deal than we thought and we have to relentlessly focus on this thing and we have to make it our number one priority for the brand in if you want to drive growth. Right. And for me that was value right there. It wasn't about efficiency, it was all about helping them realize that this was the most important segment that they should go after. Because not only does it actually win them, you know, new user acquisition, it also helps them with retention, you know, a couple of years down the road if they have a positive experience with their brand. And they hadn't thought that part through. Right.

Yogesh Chavda [00:36:39]:

So to me, you know, sometimes you have to think about what is the value that you can actually offer. Right. You can sell it as a efficiency play, but it's, that's not going to get you the longer term win, it's the value play that's going to give you the longer term play.

Adam Gray [00:36:59]:

So when a company is thinking about deploying AI and I imagine that pretty much every company is thinking about deploying AI.

Yogesh Chavda [00:37:08]:

Yeah.

Adam Gray [00:37:09]:

You know, we're told that AI can do the work of a thousand people in, you know, simultaneously and it can, it can shorten processes from months and years to days and weeks and, and everything is rosy and, and perfect. So I'm the, the, the CEO of Big Corp and I think, right, we need to be, we need to be getting into this AI journey. What, what now? Because actually it's it. Are we in a situation from many organizations where this is a solution, hunting for a problem and you know, we know that it's got all of these amazing gains to be had. But, but where, how can I use that? How can I start to make inroads with this to make my business more profitable, more solid, more innovative, whatever the outcomes I'm looking for are.

Yogesh Chavda [00:38:03]:

Yeah, yeah, great question. Right. And this is exactly the questions a lot of senior leaders are actually asking. Right. The first thing I would say is that you have an organization, especially if it's a large company, right. Or even a medium sized company, you'll have, you know, a few hundred people perhaps working in your company. Job number one, train them, give them access to the tool and train them on how to use the tool. Right.

Yogesh Chavda [00:38:27]:

Because part of the problem is that if you hold them back or you don't give them access to the full process of, you know, being able to use all the tools, then then you're just limiting their, their ability. I'll give you examples. Copilot, you know, that's one of the most popular ones at the organizational level because it's part of the Microsoft, you know, suite right now. Right. But some companies, what they'll do is okay, we'll give you access to Copilot, but we will not give you access to the, the agent building process of it. To me, that's shooting yourself in the foot because every time I have to do something, I have to go back in and re prompt the same exact thing. So, you know, what should have been an easy thing for me to do has now become a hard thing to do because I've been going back and re prompting to get that repetitive task to actually happen. Not helpful, right? So number one, give them access to the tools, educate them on how to use the tools.

Yogesh Chavda [00:39:16]:

Start with that and then give your organization the space to experiment. You know, over a six month period, 12 month period, you know, pick a time frame and say, you know, I want you, your, your department to go run five experiments and tell me what you are getting out of those five experiments. Tell me what the experiments are and how does it tie to the business that we're trying to do and let's see what actually happens, right? So create that, create a, what I call an experimentation plan, right. Coming out of the experiments, have an assessment process to say which experiments are working and which ones are not working. There was an article that came out from J and J that said that 20% of their experiments that they ran actually worked, 80% did not work. Right. Let's assume that's true for everyone, right? So if only one in five are working, right. Then at least you now know which ones are actually working.

Yogesh Chavda [00:40:06]:

And if they're working for one department, then the question becomes, how do they tie into the rest of the organization? And if there is something that's there, then how are you going to actually kind of like connect all of them together across the department so that you can actually get the output that you're looking for. In that what happens? There's going to come a realization within the company of saying, oh my God, our data is fragmented and siloed. What do we do about that? Or even more importantly, my data is actually not even clean. Right. For some people and for some companies, that becomes the barrier. And they'll say, oh, data's not clean, can't use it, therefore, let's not even do it. I've been in that situation and my point of view is you need to go into the data and start experimenting to see what actually happens. In one experience that I got, not with generative AI, but in the prediction side of it, we were using Salesforce data, and the CMO kept on saying, data is bad, data is bad, data is bad.

Yogesh Chavda [00:41:01]:

But we went into the data. 94% of the data was actually really good. That final 6% wasn't. And it was mainly because fields were missing, sales reps were not filling in the data. Sometimes it was inconsistent and sometimes there were just mistakes that happened. But at least we knew that 94% was actually accurate versus making this generic statement of saying the data is bad. Now we know. And that final 6%, you can clean it up over a quarter or two, which is what we did.

Yogesh Chavda [00:41:28]:

You have to go through this in steps, and you need to have a strategy in terms of how you want to go about doing it. Of course you need to have guardrails. So that proprietary data, PII data, all that stuff has to be protected. Give people those guardrails and then the organization will go and act against in that sandbox, if you want to call it that. But give people the space to go experiment and then give them a way of showing them how. Will you actually then start scaling it across your organization? That's kind of like what I would be proposing to companies in terms of how to go about starting the process. What I've also found in this process is that the people who are not everyone's going to embrace it right away, but there will be champions who will emerge. Give them the space to actually champion it and give them the resources to help them scale it.

Yogesh Chavda [00:42:20]:

Those who are hesitant or cautious or just saying, I don't want to do it, the cautious and the hesitance will eventually come on board. But those who are saying, no, no, no, I will never do it, then it's a question for you from an HR perspective saying that, hey, this is the new world. Are you on board or not? And if you're not on board, then what's your role within the organization? And that I think is a very legitimate conversation to have we do this with regardless of AI. We've been doing this anyways, right? So this is another use case with a new tool in this case.

Adam Gray [00:42:51]:

So have we got any really great case studies? Because Tim and I were at a conference a month or so ago and there were a series of round tables and, and the only company that I'll mention in this instance is Microsoft. They were one of the guest speakers there. But there were, there were a number of companies that were kind of consulting in the AI space, and they were tech businesses and they were Microsoft implementers and they were consulting organizations. And all of them spoke about the customer opportunity. And it became apparent. Tim and I discussed this at the end of the session. It became apparent that none of these organizations had a case. None of them.

Adam Gray [00:43:38]:

What they were doing is they were talking in either a theoretical or a. We're currently at the early stages of doing this kind of position on it. Even Microsoft, the. The only one customer case study they had was themselves Customer zero, where they were able to say, okay, this is where we have seen wins from implementing this in our own organization. And it strikes me that there's, there's as, as the, the amount of airtime AI gets and the promise of all of this change. And, and Tim and I have seen it with social, where one, I think it was a marketing director of an organization said it's not worth me getting on Twitter. It been two years ago I would have gotten it, but now it isn't. And you know, like the old Chinese proverb about the best time to plant a tree is 20 years ago.

Adam Gray [00:44:28]:

The next best time is now. So the ship for AI hasn't sailed. You know, it's not even at the point of leaving the harbor yet. So how do organizations hold on to that belief and realize that they're not falling behind, but they do need to act?

Yogesh Chavda [00:44:45]:

Yeah, no, it's a great question, you know, and quite frankly, the, you know, it's a hard question to answer because most companies, major large companies are all experimenting. There's no question about that. The challenge is that everyone's been in the experimentation phase so far. So that's why you're seeing the One off use cases that are coming out. But I would say probably in the next two years or three years, we'll actually start seeing scaled solutions that are actually having a business impact. Because we're right now in the midst of that journey right now. Because remember, this is only two years old. We're not talking about something that's been around for two decades.

Yogesh Chavda [00:45:34]:

AI has been around for multiple decades, but generative AI has only been around for a couple of years. And the accessibility is obviously, you know, the democratization of generative AI has only happened in the last couple of years. So I think we are. It's too, it's too early to tell is the answer I would give. But more importantly, linked with that, the examples of what you're seeing are actually what I pay a lot of attention to in terms of use cases. Because if one company is doing it one way, another company is doing a different way, there's something there as to why did they go down that pathway that tells you about how they think about their business or their category. And is there something to be said about that? You can almost reverse engineer and figure out what's happening. That can actually be an interesting situation as well.

Yogesh Chavda [00:46:30]:

If I may give a very quick example on what I mean by reverse engineering, just to make the point about the first power of these tools. Last year around Valentine's Day, I was talking about one of my agents that I'd built and I said, okay, it's Valentine's Day is coming up. Let me write about chocolate. So I did this, you know, exercise around, you know, ideation and innovation ideas for a chocolate brand. And I picked some company out of the uk. I didn't know the name of the company, I didn't know much about it, you know, but I said, yeah, you know, let me just pick them, you know, for fun, you know, they seem like a great company. You know, I'll do this and I'll just post it. The mistake I made when I posted it is I forgot to put a very important caveat in there, saying this is a hypothetical example.

Yogesh Chavda [00:47:11]:

I just forgot, you know, and I posted it. A week later, I got a DM on LinkedIn from some PR agency in the US saying that we talked with the, with this, the leaders of that company and they haven't worked with you. You're misrepresenting yourself. You need to pull that post out. And I was like, yeah, fair enough, you know, my bad. Totally accept that part of it. There was a second paragraph after that, which is what scared me. What Scared me was they said you expose flaws in our business model and if our customers find this out, that would be detrimental to our revenue.

Yogesh Chavda [00:47:46]:

And I'm sitting there thinking, excuse me, like what did I expose that was so detrimental to you? Like I was, I, you know, if anything, I was actually trying to help you guys for free by posting this, you know, but the learning that I got from that, first of all, you'll still find the post there. You will not find the brand name there. I turned it into Brand X basically. So I refused to take that post down because I think there's something there. But more importantly, the power of these tools can help you think through and identify opportunities. It can also help you identify weaknesses on your brand. And that can help you then do the things that you need to go do. So to your question that you're asking from a use case perspective and from a solution perspective, companies have to be thinking about the strengths of the tool in terms of what it can do for you and don't get hung up on all the things it can't do for you.

Yogesh Chavda [00:48:41]:

That's where all the airtime is being given right now. And to me that's not helpful.

Bertrand Godillot [00:48:51]:

And that would be potentially my last question. Yogesh. But the going back to the starting point of value and outcomes, are you just saying that companies are experiencing because they have no idea what the outcome can be or they are on a hunt for value or how would you say that?

Yogesh Chavda [00:49:16]:

I would say that companies have a sense of what? Well, let's put this way. The edict is go, go figure it out, right? From the CEO, let's say, or the CMO or whoever, right? It's up to middle management and the, the team leaders to say, okay, now let's go figure out how this is, how is this actually going to work for me or for, for my department, right? And that's when the value generation will actually happen. And my whole thing is that, you know, if you're going to tie one hand behind their back and then say, go figure out the value, you're not setting them up for success. So if you're going to do it, do it, go all in, do it, learn and then move on. Either the experiment is going to fail or it's going to succeed. If it succeeded, then how do I think about taking it to the next level? That should be that experimentation mindset. One other thing I'll say is that this time has, I think, forcing everybody to get out of the mindset of oh, I'll just go search and reapply and that's going to be good enough for my business. This is forcing us to actually go innovate, which means we have to be far more intentional around our critical thinking and our creative thinking that leads to solving business problems.

Yogesh Chavda [00:50:32]:

And I can actually solve some of those intractable problems in your business that you could not do without AI. But you have to be willing to suck it up and try.

Bertrand Godillot [00:50:46]:

Okay, well, thank you very much, Yogesh. This has been really great. Where can we learn more? Where can we find you?

Yogesh Chavda [00:50:55]:

You can find me on LinkedIn. Just search my name, Yogesh Chavda. If you want to use those LinkedIn things. Mine is called Yogesh AI Marketing. If you want to use that as a way to find me on Linked, I have a business page as well called Y2S Consulting. And then I have a third page on LinkedIn called Next Frontier in Insights. So on my personal page you'll find me posting practically every day about something or the other related to AI or whatever. On my business page you'll see my.

Yogesh Chavda [00:51:25]:

Well, and also on my personal page I post a newsletter every Monday morning that's starting to emphasize a lot more from a C suite perspective, what you should be thinking about. And then on my business page I'm translating that C Suite perspective into an Insights perspective as to what should the head of Insights do if you know, related to that particular topic area. So I'm actually now creating two separate newsletters, one on my personal page, on my business page. So that just started this month in fact, and then on the Next Frontier Insights. That's where I post all my podcasts that I do every two weeks. So literally in about three hours time. I'm doing my LinkedIn live session as well today, but I do that every two weeks. So you'll find all my content on LinkedIn primarily.

Yogesh Chavda [00:52:09]:

And if you want to find it on Spotify, at least a podcast, you'll find that there as well as on Substack as well.

Bertrand Godillot [00:52:14]:

Excellent. Well, thank you so much and what a great transition because believe it or not, we, we too do have a newsletter.

Yogesh Chavda [00:52:25]:

There you go.

Bertrand Godillot [00:52:26]:

So you can get the show highlights, the behind the show insights and what's coming up next in into our next episode. You may scan this QR code on screen or visit us at digitaldownload.live/newsletter your guest. On behalf of the panelists, thank you and to you. Thank you so much and to our audience and see you next time. Thank you so much.

Adam Gray [00:52:53]:

Thanks.

Tim Hughes [00:52:54]:

That's fantastic. Look at all the engagement we got as well.

#ArtificialIntelligence #MarketingStrategy #CMO #SocialSelling #DigitalSelling #SocialEnablement #LinkedInLive #Podcast