Brad Eather (00:01.569)
Hello and welcome to the Sellings Creative Podcast, a podcast exploring creativity's role in sales. I'm your host, Brad Eather, a digital communications sales enabler, helping establish businesses to sell on social. Today, we're going to tackle a subject that I've deliberately avoided on this podcast until now. One of the reasons is it's very complex and everybody's got an opinion. That's right, we're going to be talking about AI.
Gaurav (00:21.625)
Sheesh.
Brad Eather (00:31.913)
AI has made people feel uncertain about the future. It's made people excited about the future. But what I really want to unpack today is what AI actually means for us right now, because it's one thing to project what might happen in the future, but it's another thing entirely to understand how it's already changing the way that we think and operate in business today, which is exactly why I wanted to talk to someone who lives and breathes this stuff.
Gaurav (00:39.656)
Okay.
Brad Eather (01:00.767)
someone who not only understands AI, but understands it in a business context, someone with an MBA and a background in media communications and AI focused data science. Please welcome to the show the head of AI strategy and solutions at Warp Development, Gaurav Devsama. I hope I got that right.
Gaurav (01:22.654)
That is correct, Brad. Thanks for that. Thanks for the warm intro. And I'm really excited to be here. As you said, I think it's a space where
Everything is changing so fast. And I think it's always interesting to talk about different verticals in that sense, or jobs, and how kind of AI impacts that. So very keen to get started today.
Brad Eather (01:47.819)
So like you said, AI is changing at a rapid rate, like so rapid that it's impossible for everybody to understand every single aspect of AI and where's it going? What kinds of businesses do you focus on at Watt Development?
Gaurav (01:48.814)
you
Gaurav (01:58.829)
Okay.
Gaurav (02:04.428)
Yeah, at Warp, we focus on pretty much everything really, in terms of we don't look at niches in that sense. And that's one of the main things about AI, because it's affecting everyone. It's not something that is just for one industry or another. Obviously, we do see patterns where the immediate return of investment is different for different industries. For example, I think
Manufacturing is a very interesting space in AI because historically, a lot of manufacturers did not actually make the jump from, let's say, step one to step two, three in terms of things like having a CRM or having those kind of sales products that they can use like an ERP system, whereas it's a lot of just actual spreadsheets where complex modeling for pricing or operational data is just set up.
And quite weirdly, now AI actually helps them jump from, let's say, step one to step four, because suddenly you don't need to actually hire someone who's an expert in managing Salesforce, for example, or HubSpot, because what you can do is you can build the solutions and create custom software for these firms based on the data they already have. Obviously, the quality of data and things like that, which I think we'll talk about later.
Today, I'll go more deeper into that that matters. But I think you can see the shift in that one industry where suddenly you can bypass some of the steps that you'd normally kind of be forced to take in a non AI world. So that's one. I guess with also manufacturing, I think there's a lot of issues around knowledge retention. And that's probably not just manufacturing. I think it's across all industries where
A lot of people would have stayed in the business for 20, 30 years. And now they are coming close to retiring or just leaving the business. And then what happens normally is you lose all that knowledge that they have accumulated in the business because a lot of it is just in their head and it's undocumented and things like that. So I think now we have it, we are kind of changing this where we are trying to quantify as much as we can, all of that. So, you you're bringing this knowledge out from their head in a way where
Gaurav (04:26.028)
you create documents and you create like some sort of, I guess, any documentation that can be fed into the AI. So even after people leave the business, that knowledge stays. And then I think that becomes almost your IP. So yeah, I think we can chat about all that as we go forward today.
Brad Eather (04:36.876)
Hmm.
Brad Eather (04:44.577)
Perfect. So before we go any further, I want to sort of kick things off talking about a little bit about you. And at the top of the interview, I mentioned that you've got several degrees from communications to data science. How do you think that that knowledge and experience has given you a unique perspective on how AI is playing out and how's it set you up to where you are now?
Gaurav (05:11.774)
Sure, and that's a very interesting question because I think when you look at people who are in AI, I don't really have a very conventional background in that sense. And I think that's the advantage I have is I look at it from a business perspective rather than I think it's very easy to get lost in all the technical details. And I think that's been my advantage coming from a diverse background in terms of I'm looking at this from a perspective of
Okay, I forget about the AI. What is the problem out here or what is the opportunity out here? And then how are we going to actually solve it? we need AI or is there something else that we can actually do? And then you go backwards from that and then use AI if it's needed. I think what we see a lot in terms of some very talented individuals who are very heavily, you know, they would have done PhDs and they have a very heavy data science and AI background.
They focus on all the technical details and they're incredible in that in terms of their understanding of, let's say the statistics, the math, and also the understanding of models and things like that. But the thing is when you are solving a business problem, the stakeholders don't care what models you use or what math you are kind of, I guess, preparing to solve something. They care about, I have this problem, this keeps me up at night. How can you solve it?
And I think my advantage really has been that with my background, I focused on that just quite naturally rather than looking at all the technical aspects and the coding aspects, which I think would be the case if someone's heavily coming from an AI or machine learning background.
Brad Eather (06:59.317)
Yeah. Cause I think just for the audience that maybe, you know, AI is confusing for quite a lot of people and you sort of bought up two different aspects. think that a lot of people are using AI chat GPT, know, that's, that's what they're using. And I think that there's sort of two ways that fundamentally you can use it. One of them you've bought up models and, we're going to get into AI agents. But I think that there's.
One way that you can use it, which is using ChatGPT and you're using it for creative output or a partner in writing and synthesizing your ideas, getting words onto a piece of paper. The other one is more like models. So more like a knowledge hub, if you're not familiar with the concept of models. And essentially what that is, it gives you access to large process.
large amounts of information. But I think that where AI has taken the leap and correct me if I'm wrong, but it's in qualitative data. What we used to be able to do with quantitative data is find, find commonalities. AI is now giving us the ability to find commonalities within qualitative data, which is much, more rich. So I want to hear your perspective on those two models.
Gaurav (08:14.493)
Okay.
Brad Eather (08:24.631)
first of all, from maybe put yourself in the position of a salesperson and maybe how they could use it, the creative side, just generalized AI. And then let's put it into a business context and talk about AI agents and how we can actually start leveraging IP for business advantage.
Gaurav (08:49.87)
Absolutely, and those are great points. I think when I answer your first question, I might actually annoy a few people to be fair, because the thing is, I think, yes, it's good to use tools like ChatGPT for, let's say, outreaches, be it on LinkedIn, be it on
Brad Eather (08:59.563)
Mm-hmm.
Gaurav (09:11.766)
I guess, through emails and things like that. But that's just, I would say, a starting point. And I think that's, yeah, you get the intuition behind that in what works, what doesn't. But where I think I am more excited about is making sure a salesperson, because traditionally, most, let's say, if you're not in technical sales, salespeople wouldn't have a lot of the technical understanding if you're selling a software product.
for example, whereas I think with AI tools now, it pushes salespeople towards those kind of technical understanding. And I think that's the thing that I am more interested in rather than creating personalized outreaches and things like that. Whereas I think it's about if I am trying to create a proposal, for example, for some software implementation that my company is selling.
and I don't really have a technical background with chat, GPT and other tools. Now what I can do is I can actually create a proposal, maybe not a hundred percent, but at least I can get pretty far with some of these technical details. If I look at my other projects, what we have done as a company and then get some like understanding from chat, GPT or other tools and really kind of.
I guess, save time between going back and forth with, let's say, a developer who would help me with the technical things, technical aspects in the proposal. And you can be quite creative because then you have your capability of sales. And then you're also adding AI tools as capability of understanding some technical stuff and then saying, you know what this is?
what we spoke about in the meeting with a client. This is the thing that the client wants to achieve. These are their problems. How can we make sure that we create a really good proposal? And that's the starting point again. And I think never, at least yet, just blindly copy and paste stuff. It's a good starting point, but then you also have to put your own mind towards it. So that, I think, is quite valuable, in my opinion.
Gaurav (11:28.142)
But I personalized outreach is actually something that I just think there's so many companies that are doing it right now in terms of just scraping LinkedIn, scraping other social media, and then sending personalized emails or LinkedIn messages. And I just think it's pretty much a race to the bottom. To me, it feels like one of those tar pit ideas where initially it looks good, but now there's like...
500 companies every day that rock up doing the exact same thing and getting funded by a lot of them are funded by let's say Y Combinator, which I definitely find a bit surprising. But I think what that ends up happening is you have this absolute slop of emails in your inbox and it can be the best personalized email, but if I have 200 emails every day coming,
into my inbox. I think that becomes a problem no matter what. obviously there's email filters and things like that, but I think still that's not what I think would be valuable in the future for like AI in sales, for example. I know I've spoken a fair bit out here, so you can feel free to ask me any questions.
Brad Eather (12:42.805)
Yeah, so maybe let's let's hold off on going into the agent discussion for a second. Because I think that what gets lost in this conversation of AI is the fact that there's a human is this is this is this is just as much about humans as it is about AI. And I know that maybe you both share the commonality in the fact that we've both done a communications degree.
Gaurav (13:03.214)
Mm.
Brad Eather (13:10.847)
And there was a philosopher that we're both familiar with called Marshall McLuhan, who wrote the media is the message. And one of his core philosophies was that as technology develops, any advancement in technology can be seen as an extension of ourselves. One of the easiest ways to understand this is just jumping in a car. Without a car, a car gives you the ability to go further.
but you can also feel the car. You become one with a car. can feel where the edges are, for example. But AI is doing that. It's changing a different aspect of what it means to be human. But what I wanna talk about is sort of the behavioral changes, the traits and things that are changing. And I suppose going back to...
Gaurav (14:02.99)
Hmm.
Brad Eather (14:08.011)
going back to what we were just discussing is like.
weird as if you were taking that example of the email, right? And all of a sudden, we're oversaturated by the emails, the human becomes desensitized to that.
Where do you think that...
How do you think we should be using AI to actually stand out and be more human, be the author of AI and using it as a actual tool to help us?
Brad Eather (14:44.907)
be more unique than just another AI agent talking to another AI agent, suppose. What's your thoughts on that? I'm not sure if questions, if the questions all that clear.
Gaurav (14:54.081)
Sure. No, that was really good. That was really good. think.
I'll give you an example, I think, which I find really valuable is I will, you know, some of the AI voice things so you can talk in real time. So what I'll do is I'll give context about the company and a meeting that I might have, an important meeting. And then I will actually simulate that meeting with the AI, let's say inside ChatGPT.
Brad Eather (15:09.697)
Mm-hmm.
Gaurav (15:28.544)
It's absolutely incredible and kind of creepy how many times the AI has actually come back and asked me questions, which I thought, that's a unique question. And it's the same question I get asked in the meeting later. So it's that kind of preparation in that sense where it's like having access to an expert 24 seven in, let's say sales or any, any other kind of domain. So that's.
Brad Eather (15:40.182)
Mm-hmm.
Gaurav (15:56.576)
I will say, I think I'm a bit biased in this because I think naturally, like I always been dumb enough to kind of start doing things that I have no idea about. but I think with AI, it just takes it to the next level because now I might not have an idea about something, but I always know that I have this kind of copilot or like an intern that'll kind of give me info about. So that's a big kind of behavioral change.
Brad Eather (16:07.702)
Yeah.
Gaurav (16:25.326)
I feel if anyone that uses AI, let's say daily in their workflows, they would kind of, I think, relate to this where it's like, you can say yes to a lot more things because you have, if you have that agency, if you have that initiative, you just know that you'll kind of figure out things that you might not have any idea about. And that changes a lot, right? Because it's like, if I have a sales meeting and it's highly technical, I don't have a developer available today. And that happened a lot when I was...
Brad Eather (16:51.511)
Hmm.
Gaurav (16:55.566)
working at the data consulting firm. It was always like, because developers have their own kind of job to do, so they might not be always available to jump on a call with you if you're non-technical. But now you can actually still prepare quite well for those meetings and then you can immediately see how much of a behavioral change that is. You can create demos now as a salesperson, you can create demos of something that the client might want without writing a single line of code. So again,
That's another kind of big change. I don't think it's happened broadly yet, but I think the initial aspects that at least we are doing a lot at warp kind of show where the world's kind of heading.
Brad Eather (17:36.535)
Yeah, because something that I believe, especially from a sales perspective is that soft skills is where I would be putting, if I was a sales person coming into the job for the first time, I'd be saying, develop your soft skills, develop your soft skills. Because you said something really interesting there. You said, you're dumb enough to try things that...
that you haven't tried before. But what I think that AI actually does, gives people, rather than niching down, it gives people a lot more contextualized information. So if I'm dumb enough to go out there and find the information, contextualize complex information, and dumb it down in a way that somebody else can understand, that as a salesperson is far more beneficial.
Gaurav (18:19.755)
it
Brad Eather (18:33.074)
than having the technical know-how, just having that broad understanding. How do I get the information? How do I understand that information? How do I communicate that information? Like it really makes it a lot simpler. Whereas you might have to spend hours with a developer trying to understand the nuances of what's possible. Now,
Gaurav (18:45.39)
Hmm.
Brad Eather (18:58.199)
we can just get the, we can literally ask exactly what we want and just get the dot points. One of the hardest parts about getting information out of, a seller and a developer are two very different people. And sometimes sitting down and having a discussion with them, those two people, you could spend hours trying to get the information out just because you're not talking on the same plane. You're asking the questions and their ability to answer it the way that you want it answered.
Gaurav (19:08.014)
Hmm.
Gaurav (19:16.849)
Yeah.
Gaurav (19:22.286)
Mm.
Brad Eather (19:27.998)
is just
It could take hours. It could take hours. And I really like the idea. I personally love how AI is just being able to simplify the information. If it's not the information that I want, I can then just reiterate. And I mean, if I can't get the answer within three prompts, like I'd be very, surprised.
Gaurav (19:32.703)
Yeah.
Gaurav (19:47.79)
.
Gaurav (19:52.128)
Yeah, I think 100%. And the other thing I think that is, I guess, happening is, I think we are moving from pitching to co-creation in terms of a lot of these solutions, because it's what will end up happening. And I'm talking about this to a friend, actually. It's everyone's got their note takers now.
Brad Eather (19:52.673)
Sir.
Brad Eather (20:06.849)
Mm-hmm.
Gaurav (20:21.678)
that are doing real-time transcription and stuff like that in the meetings. So as a salesperson, I'm in a meeting with a client and my note takers there, my note takers transcribing in real time on what the client requirements are, what they're doing. I can just connect it to a designing tool that's run by an AI and that tool can then create designs.
and demos like in real time in terms of, and that becomes, think, like, you're not just pitching, you're actually bringing the client to create something with you, right, in that goal, or even if it's not in that goal, but later. But I think that's, I feel like it's, I still don't, I don't think I'm explaining it too well because I think it's still kind of fresh in my head.
Brad Eather (21:11.223)
Let's explain what an agent is. Because I think once we've explained what an agent is, this chat will make a lot more sense. So tell us, what's an agent?
Gaurav (21:18.126)
it.
Gaurav (21:23.224)
Sure. That is a very interesting question because if you ask 10 people what is an agent, you'll get 10 different answers right now. So in my definition of an agent, because it's very easy to get lost between automation versus an agent, so I'll start with automation where I think normal traditional automation is basically where you know exactly you can predict what's happening. So it's a lot of like,
kind of boxed assumptions where there's only a limited number of things that can happen so that you can kind of build a solution to just focus on that. Whereas now with an agent, you don't need to code it in a way where it's only like, if this happens, that happens, that happens. You don't need to do that. Whereas an agent can decide on the go, on the fly, based on what's happening, that, okay, I need to do X or I need to do Y.
So what we are doing, that's, you can do either a single agent, which is basically you're connecting this AI models to carry out a task for you. For example, if, sorry, I'm just trying to think, okay, update something in HubSpot based on a call I had or something like that, right? And I think that could be an agent, could be just normal automation based on what
I guess the exact use case is. So that's one agent. Basically, you can create these solutions that can make their own decisions on the go based on what's the best criteria. And you can also create multi-agents, which is one agent, let's say, the research for you for a prospect. And then it provides that research to the second agent, which then creates a proposal
based on whatever's been done and what your company information is. And then the third agent might actually kind of based on that generate a lead score on how likely this prospect is going to convert or something like that. those are, yeah.
Brad Eather (23:33.975)
So, cause my, this is my understanding for anyone. Like if we bring it all the way back out to an LLM, like chat GPT, what an agent does is contextualize information. So we set up parameters around what information we want it to have access to what information we don't and a potential task that we want it to fulfill. Is that right?
Gaurav (24:02.862)
Yes, basically, think agents are, they move from passive AI, which is let's say, charge GPT, I would call it as passive AI, because you are asking something is providing you a response, but it's not only doing something. Whereas with agents, it carries out a task or set of tasks. So that would be like my kind of definition. But then again, I think depending on the definition, guess you can.
Brad Eather (24:19.734)
Mm-hmm.
Gaurav (24:31.916)
I get dissected different ways, but yeah.
Brad Eather (24:32.887)
So, like if we were, I think everybody will be familiar with like Microsoft Copilot in terms of a meeting. Most people will have access to transcript notes and that will, it'll give you the meeting notes for the discussion that you just had. That is an agent in itself, right?
that is an AI that's been set up for a particular set of purposes. And the outcome is at the end of the call, you've got a transcript and you've got your meeting notes. Very basic. But in a business context, what we're talking about now is like every piece of IP that a business may have related to anything from the sales funnel to the technical aspects of a job to...
Gaurav (24:59.148)
Yeah.
Brad Eather (25:25.419)
the legal requirements that you might need to be familiar with. We can actually create an AI agent that is essentially like a, a master employee, suppose, someone who's had.
50 years of experience throughout the entire lifespan of the company and pass that information from one agent to another to another to get the outcome that we want. And that's what we're working towards in the area of agents. Yeah.
Gaurav (25:56.216)
Yeah, that's, guess you don't always need to hook into your own data.
in terms of the knowledge base, but obviously for any business, that's where the value is. It's connecting your SharePoint to an LLM, obviously within secure environments like Azure or AWS, so your data doesn't actually go to third parties and it's all protected. Yeah, because I think that's one of the main questions I get all the time, it's what's happening to the security and cybersecurity side of things.
Brad Eather (26:20.235)
Right, point.
Gaurav (26:29.268)
And I think it's all protected for when you host it inside like the Microsoft and Amazon services. So that's not an issue at all. So yeah, I think that's one. And also some businesses would like to connect it to their Salesforce data, HubSpot data, and then fetch info from there. So basically, it's kind of crazy because it's such an interesting space. What's happening is over the last 20 years,
Brad Eather (26:44.215)
Mm-hmm.
Gaurav (26:58.286)
everything's got like SaaSified. It's like, there's a SaaS for this, there's a SaaS for that. And now, enterprises have the data in like 20 different systems. But for this AI agents to do the task well, and also to unlock new capabilities, what you need is all the info in a central repository, which is basically a lot of what we do. Because, and then what happens is then you can really mix and match those qualitative data.
Brad Eather (27:21.099)
Mm-hmm.
Gaurav (27:26.478)
that's scattered over everything and start to, know, merge them to find those insights and patterns and really generate those kind of AI related information that was previously just not available.
Brad Eather (27:40.503)
So is it fair to say that if there's a small business out there who's considering it, part of the element of business change in this environment is organizing information, existing information in a way that AI understands.
Gaurav (27:57.654)
That is definitely one of the most important points, because it's like garbage in, garbage out that has happened in traditional data modeling. That doesn't really change now. the other thing I think, that the models are getting quite strong to even pick up some of the structures where it might not be optimal. And I think as this keeps getting better, maybe that'll improve. But it's always a good
practice to make sure you have naming conventions, because a lot of the times, know, you, to be fair, most organizations would just have a document one, that's it. And then you're thinking what is document one? and then, or things like that, or you will have, 20 versions of the same document. And then, different people are making updates to different versions of the document. So you don't really know which one's the last kind of main document that you should look at. So things like that will confuse the AI when you build solutions.
So it's kind of important to have some sort of like data management policies on your organization. That's always a good place to start. And the other thing I I tell everyone is start thinking of creating documentation, creating files, not for humans, for like AIs, because in kind of a machine friendly format.
And it's like, so it's like, for example, for markdown files, like dot MD files work really well with, you know, giving instructions to AI and things like that. So, and the other thing is if I have a complex Excel spreadsheet, traditionally, I won't actually, you know, state out maybe all the formulas that I've used or why I have done something what's like in just text, right?
Brad Eather (29:23.799)
What does that mean? What does that mean?
Gaurav (29:52.974)
But now I would actually suggest people, when they're building these models and stuff in Excel or wherever, write down in text on what you have done, what's the value behind it, what's the reasoning behind it. Because then when you feed it into an AI model for analysis, the AI has more context. So you are actually creating this stuff for downstream AI solutions. And that's like a behavioral change, which obviously no one would have done before. Or you are having...
and to create documentation and pages and pages of documentation to just instruct the agents on what they should do when they. OK, if you look at this file, then always do this. Or if you you go towards this share SharePoint site, you should do this. So all of that's becoming instructions.
Brad Eather (30:40.471)
So what I'm hearing is like contextual information, but I think what people get misunderstand is that AI is very good right now at analyzing certain types of information. So if we had a look and said AI could read, AI can see, and AI can hear, right?
If we were to put it down to that. AI can read very, very well. And I mean, until last week, maybe my opinion has changed this week, but AI couldn't see as well. You know, you can't put, if you were to put an image in front of it, its ability to understand that and contextualize an image is different from if it could read. Yeah.
Gaurav (31:32.81)
Depends. think the vision abilities are actually quite good. And I am always surprised that there's a feature in chat GPT, for example, where you can switch on the camera and in real time you can show the camera around and the AI actually looks at what's happening. And I have been amazed at how accurate it is to actually tell me in real time on what is around me or things like that, which is, and I think it's a feature that I thought would be a lot more
would have a lot more hype, but it hasn't because for me it's immediately it's like, if I'm a manufacturer and I work with a technician, for example, in a facility, that would be incredible for me because if I can just show it around and then it connects to my database or SharePoint with all the info, then immediately I have like a senior engineer who can help me out without me having to actually go out and ask someone, you know what? There's a fault out here. What should I do?
and things like that. I think its vision capability is actually really good and it's only going to get better.
Brad Eather (32:37.633)
Yeah, so what would you be advising your clients? What format would you, if they were building an agent, if it was just, say for example, they've got a bunch of video content, they've got a bunch of written content, where would you start in that feeding it into an agent?
Gaurav (32:56.686)
What do you mean by format?
Brad Eather (33:00.585)
Content format so yeah, yeah
Gaurav (33:02.996)
OK, like videos, basically that. I think that definitely depends on the end use case. So I think depending on what's at play, I think that's where it's kind of hard to give, I guess, one answer in terms of that. So I'm not 100 % sure. But yeah, basically, I think it could be anything because these things are so flexible. It just depends on what you're trying to achieve.
Because for example, one of the most popular use cases is looking at RFQs and then the AI analyzes RFQs, is a bit of text analysis plus vision, because there's some, there'll be drawings and things like that, need vision capabilities. And then it'll go back and look into your past projects. What profit margins have you made with similar projects and things like that? And then marry up the RFQ analysis with that.
And then you can build different agents for downstream processing, which then will give you the actual code or estimates on what this project, I guess, might cost you. And you can have vision capabilities. You can have text capabilities. I think those two are probably the most popular right now because we haven't really gone to the audio stuff yet because most business data is not really in audio format.
Brad Eather (34:28.011)
Yeah. Yeah, right. So I, I know at the beginning of the podcast, I mentioned that I wanted to talk about I wanted to avoid the future. But let's go let's let's go into a bit of futurist thinking. Because they're it's hard to talk about AI to talk without talking about the implications of maybe where it's going. But I want to talk about hiring rather than firing today. And I wanted
Gaurav (34:36.476)
Okay.
Gaurav (34:40.558)
you
Brad Eather (34:57.953)
think about a question that I have, and I know that you probably don't have the answer, so it's just worthwhile having a discussion about it. pathways, if now we have the ability to make essentially agents with high level thinking, where does the role of a junior coming straight out of university or school fit into the context of all this now, when there's just such a massive jump of fundamental understanding of some of
that, know, whatever the role it is that they're in to actually applying it. What do you think, where do think the pathways of a junior coming into the professional workforce might end up?
Gaurav (35:31.886)
you
Great question. First, think, you mean junior in one specific domain or do you mean just in general?
Brad Eather (35:48.265)
In general, so yeah, in general, if it sounds like you've got quite a good understanding of the manufacturing space, so maybe like even just an engineer coming into the manufacturing space.
Gaurav (36:01.944)
Sure, I think this is like one of the important almost dilemmas. And I think you'll see on LinkedIn and on Twitter about software engineers probably being the most discussed topic in terms of yeah, what happens to junior software engineers, for example, because if chat GPT is at the level of like a junior engineer in terms of coding abilities, and it's only going to get better what happens. For me, I think the roles change.
That's what happens in terms of, and obviously I could be absolutely wrong out here. Making predictions in the AI space is the worst thing you can do because everything changes so quickly. But for me, think it's about, because to create this agents, you will need a lot of human input in terms of what you want out of it, right? In terms of documentation, preparing the architecture and things like that.
Brad Eather (36:39.999)
Yeah, yeah, yeah.
Gaurav (37:00.718)
And depending on the complexity, maybe one of the roles that juniors will have is basically doing that, where you're kind of maybe looking at what prompts work better for like what instead of like writing, I guess that's like the first stage of someone writes a set of prompts and then, or like set of instructions, then that gets passed to a senior developer who then kind of fine tunes on what works, what doesn't. if you look at it, it's a bit like,
writing code right now, because you start with some basic stuff and then if you see any developer would look at it and change stuff and fix bugs and things like that. So it's similar in terms of that. And I think it's also about, I think the senior devs in this scenario might be more about managing agents where it's almost like what I like to call it as HR.
for AI agents because there'll be so many AI agents running around in organizations. Someone has to manage them, be it an outsource provider or be it like a senior dev. So look at performance, look at accuracy, things like that. But the initial aspect in terms of designing an agent and bringing those kind of different, almost blocks of Lego together probably could be something that is done by a junior. And I think, and I also feel like...
I feel like more and more of these roles will end up getting pushed towards that kind generalist flavor where I think right now if there's a junior programmer, the junior programmer just writes code pretty much. Whereas I think maybe they're involved more in sales meetings where they're kind of drafting, designing, creating demos in the sales because they're not sure.
Brad Eather (38:35.767)
Mmm.
Gaurav (38:55.264)
but I think roles are definitely going to change a lot.
Brad Eather (38:59.617)
So, cause what, just thinking about, I mean, the last major shift that we had was maybe social media, right? And I'm marketers coming in, all of a sudden junior roles became social media marketers. then...
just having a discussion, right? So the junior roles become AI practitioners in this example. All of a sudden we have a massive gap in understanding. And I think that part of the challenges that business faces with social media is that
It's devalued work. There's such a high level of understanding that the people at the top who don't understand what actually goes into it and the nuances become devalued. And it's, it's the ones it's the people who are practitioners who actually have been there and done that and understand the nuances who actually become successful on a social media platform. I wonder if, and we're, just talking.
Gaurav (40:03.48)
Yeah.
Brad Eather (40:04.011)
We're talking futurist thinking, no one's, no one's predicting anything here, but I'm wondering, I'm wondering if the same sort of thing plays out in, the role of new AI when, where, there's a gap, there's, there's almost a generational gap in digital understanding of how these things work in practice. And I think that the reason that I think it's important is because that if that's devalued,
Gaurav (40:07.335)
Yeah.
Brad Eather (40:32.299)
the opinion of the people doing it is potentially devalued as well. When in actual fact, maybe they should have a place at the table of decision-making. Does that make sense? When I'm talking about that gap, there's a gap. It's like, it's a junior role, but the knowledge that they have is actually influencing, should be influencing decisions.
Gaurav (40:55.508)
But I think good organizations will not have the gap.
Brad Eather (41:01.195)
Mm.
Gaurav (41:02.846)
Like I would say, it's because in any organization that kind of is well structured, if you are someone who's actually, you you have a lot of agency and you're coming up with good stuff, then I think that gap shouldn't exist. And in organizations where that gap does exist, I think that's no different than what is now. I would say in terms of, think it would be a similar gap that exists now in, let's say, non-AI context. So I think that's...
Brad Eather (41:24.82)
Yeah.
Gaurav (41:31.914)
I would say that's not even...
Brad Eather (41:35.115)
Yeah. Yeah. I mean, I tend to agree. The thing that I would say that is that maybe AI is more pressing of an issue. Companies that have that gap. Maybe they need to address that gap pretty quick in this scenario where they've had 20 years almost to deal with social media. I don't think they're going to get 20 years to adapt to this.
Gaurav (41:43.126)
Mm.
Gaurav (41:57.838)
But I think if you are, let's say a junior employer, if you're in uni, this is such an amazing opportunity for you because if you start using these tools, and then you come onto a meeting or you reach out to someone senior in your organization, or even forget about, let's say you don't even have a job, you're coming to an interview and you research the company and then you build a prototype that they might want.
Brad Eather (42:05.1)
Yeah.
Gaurav (42:26.622)
using AI tools, which had previously not been possible. Suddenly, the company has to listen to you. They're forced to listen to you because I think a lot of these stakeholders, don't really have the awareness yet. So if you are someone young, you are a junior and you have that initiative, you can actually blow their minds. then that kind of messaging gap that you're talking about doesn't really exist because there is no other way but to take you seriously.
because it is such a compelling factor. So I think that's why I tell everyone that play around with these tools. then the only thing that honestly stops anyone to build really cool stuff right now is really just themselves, because there's so much that you can do and you can test out the boundaries. For example, let's say I am a sales intern and I see that the head of sales at this company
has a massive problem in terms of, don't know, prospecting or something. Just throwing out an example. But right now, even without a lot of coding abilities, I can kind of patch together something that kind of barely works to maybe show my initiative that, you know what? This might help you. Then suddenly you are going like four steps ahead to the head of sales to say, you know, I tried building this for you.
because I think this might help you solve the problem. It might not help, but it starts that conversation. It enables someone to actually move way further than what they would have done earlier.
Brad Eather (44:06.743)
That's a call to action for anybody that's watching this at uni right now, for sure. And I think that that...
What we're talking about here is a very different way of thinking. And I think that in the context of this podcast, Selling's Creative, it's a great way to come into the last question and ask you with everything that you know, what's your definition of creativity?
Gaurav (44:37.61)
It's a really interesting question. And I think I always love when someone asks me this because if you look at traditional creativity, like in terms of artistic abilities and things like that, I have none of it. So for me, the definition of creativity is a bit different. How I look at it is as being really creative because this is not just AI, this is just the internet in general. There's so much info.
We live in an age where it's knowledge is just abundant everywhere. And I think it's the best time to be alive in that sense to kind of just soak up all of that. And I think where I see creativity is how can I use this kind of odd shaped Lego pieces to kind of bring it together to actually create my own version of something. And I think that's where I feel like I have my strengths and where I creativity fits in for me. It's looking at things that are
a bit different. It's like trying to see, you know what, this piece of Lego doesn't actually fit in here, but I feel like there's something that I can do. Maybe shave off a bit on this thing to kind of merge it towards this and actually create something that solves a problem that no one actually looked at earlier. So I think that kind of creativity is what I think I bring in.
in terms of just making sense of this kind of different moving pieces and then almost bringing them together in kind of a non-conventional manner, which has served me well because in an area where there's like so much chaos right now in AI, I think if you are just looking at it from a common sense or a conventional side of things, you'll get overwhelmed quite quickly. It's all about bringing
these different parts together and then kind of building stuff that actually solves something. I hope that answers your question.
Brad Eather (46:37.495)
Yeah, and I'll bring it back to that answer that you had before regarding how like, you know, maybe applying for a new role because I really like that. You essentially turned the the interview process or the application process upside down on its head. And you said, all right, if AI is possible, I can build something the purpose of it's not to be perfect. The purpose of it is to show initiative. And I think
Gaurav (47:04.855)
Hmm.
Brad Eather (47:06.901)
That is a really creative way to get the attention and show initiative in the job application process. And I think that what you've just described, your definition of creativity is really flipping that thinking on its head and adding, adding something, shaving those little Lego pieces and fitting them into to suit your purpose. I like it.
Gaurav (47:30.082)
Yeah, 100%. And I think that AI definitely helps you to be really weird. And I mean that in a good way, in terms of how creative and how you can just go and do things, I guess. And I think that's been my experience. It's like, I've got this tool that
Brad Eather (47:40.412)
Mm-hmm. Yeah.
Gaurav (47:56.396)
I can do anything with pretty much and we are not like so early to this and I don't even know what this will actually look like in two years time because it's exciting and also scary in terms of the capabilities that might come.
Brad Eather (48:10.529)
Yeah, absolutely. Well, Guruv, it's been a pleasure having you on the podcast. Where can we find you if you're looking for you?
Gaurav (48:19.896)
Sure, you can find me on LinkedIn. My name is Gaurav Dev Sharma. So I think you can put my LinkedIn profile somewhere in your podcast links. And also feel free to reach out to me at gaurav at warpdevelopment.com. If you have any questions, that's my email. And I'm always happy to chat about anything and everything that's AI related. Like I love nerding out just kind of anything that.
It's easily because this is such an exciting space and I think I am at times kind of start off of chats about just nerding out AI. So all this happened to kind of, if you're an athlete, grab a coffee or otherwise just reach out to me on LinkedIn or like on email.
Brad Eather (49:03.031)
Perfect. Well, everybody out there, thanks for listening to the Selling's Creative Podcast. If you've enjoyed this episode, remember to subscribe for more. And in the meantime, happy selling.
