Rolando Rosas 3:27
start let’s start there. Because how do you what inspired you to be a data scientist?
Julian Ghadially 3:33
I fell into it. I fell into it after college. I got this data science degree from University of Texas, and I instantly saw, Wow, there is so much coming down the pipeline. This is, I think this is 2015 2016 I went and did strategy consulting, but always had that data science skill in my back pocket. And then when the AI wave hit with ChatGPT being released. I realized this is the time to be an entrepreneur in AI. There is, there are so many jobs out there that can be enhanced with AI. And if you’re passionate about entrepreneurship, you cannot miss out on this opportunity. You know, there’s
Rolando Rosas 4:16
a lot going on. And now the when you when we talk an opportunity, you know, the probably the quintessential guy right now when it comes to entrepreneurship, and entrepreneur and entrepreneurs would be Shark Tank favorite, Mark Cuban. I want you to listen to what Mark Cuban had to say about AI. Or go ahead and roll that. How do you see this new AI era.
Mark Cuban 4:42
It’s a beast. There’s only two types of companies in this world, Morgan, those who are great at AI, and everybody else. If you don’t know AI, you are going to fail Period. End of story. Whether you are an employee, you’re going to have to understand it impact, how it impacts your job, or how you can use it to be better at your job. Sam, if you’re a student, the same thing, and if you’re a CEO, you can’t just say, Okay, I’m gonna get my my tech guys to understand it and educate me on it. You have to understand it, because it will have significant impact on every single thing that you do. There’s no avoiding it.
Rolando Rosas 5:16
There’s no avoiding it. What do you think about what Mark just said,
Julian Ghadially 5:20
I love that. I fundamentally believe that in the future, the AI is going to be smarter than humans. At least in some domains of intelligence. There’s nothing scientific that says that humans have to have the smartest brains out there. At the same time, as a realist and as someone who’s been tinkering around with different opportunities. I do see a lot of industries that are very relationship based. And for for those, for those businesses, I think those might be, you know, some of the later ones to be innovated. But absolutely you cannot ignore it.
Rolando Rosas 5:56
You know, there’s no getting around it. And one of the I had a guest on last week Neil Patel. He’s when it comes to SEO, marketing and content. He’s probably one of the leaders in the forefront of that, and he was sharing that when it comes to blogs and traffic, SEO traffic, that right now, the human written content for blogs leads the AI written content, content by about 5x and I know that you, you, you’ve done some also some analysis and some studying in this and that there are some other things, like product based ads from an MIT study that you had, you had mentioned some time ago that they found that the AI content, are we looking at different things and aspects when it comes to blogs versus ads, versus other aspects of work where AI may either assist humans or make humans better? Or are we just very early, and there’s just data all over the place that will support different things about how to best use AI.
Julian Ghadially 7:01
I love this question. And also, when it comes to studies, it’s all about how you do the study these there is no final answer. This is constantly evolving there. The MIT study you bring up is very interesting. They they had a top advertising firm create product descriptions for advertisements, and they gave it to 1000s of readers out there to score the satisfaction with the content and the willingness to pay. What they found was the AI did, it had a higher satisfaction amongst the readers and a higher willingness to pay for this, these product advertising content that is much different than this blog finding that you have. And I think with the blog finding, I’ve tried to do some AI generated blog content. If you just ask app ChatGPT to write your blog for you, no, it does a much worse job than if you get an AI agent to write your blog for you. With agents, you can have one AI that goes out and does the research. Let me
Rolando Rosas 7:59
ask you something about that. For those that are not familiar with an agent, what’s the difference here? And explain to us what an agent is when we’re talking about AI,
Julian Ghadially 8:09
it’s a difference between an individual and a team. Two brains are better than one. With an agent, you put multiple GPT in a in a room together to have a conversation, and you give them different roles. You give them access to different tools. You can have a research GPT that goes and scans the web, looks at other blog posts, look at looks at textbooks and gathers information, provides it to a writer that then drafts the blog. And you even have an editor GPT that takes what the writer wrote and provides feedback on style, content, et cetera. That does way better than if you had one person writing or one AI writing in a room by themselves.
Rolando Rosas 8:50
And so all of this these different roles. So let’s call it roles, right? That an AI can do. You’re saying, you know, you have a role for an AI that’s doing research. Maybe that hands it off to a writer style role, and then you have another, an editor that maybe finalizes that and looks at the work from the other two. And then a human would be part of that team to see that it all makes sense. Is that what you’re saying,
Julian Ghadially 9:14
I love that you bring the human in there. Usually we think of the human as coming in at the end. Making sure it all makes sense. With this MIT study, they tried both the approach of, let’s have the human be the editor, or what if the AI is the one that has the final call, and when the AI had the final call as the editor, it actually did better for the product, advertising, content. Wow. Like to me, that was surprising. I think, I think what the reason why the AI did so well as an editor is AI can write very concisely. It can get to the point human editors. I don’t know we have our bias toward going down some rabbit hole, but I tried this out with some Google ads. I launched my own ads, and I, I was, I was the writer. I was the editor, my headlines, my descriptions were were very much focused on my company, my value proposition, my offering. I liked the AI suggestions better. They were focused on the keywords, the customer, what other competitors were having as well performing ads. So the AI is not always the one that’s going to be generating the content. Sometimes it is the approval and it’s the editor at the end of the flow. Interesting. So
Rolando Rosas 10:30
what are your thoughts about that being used? So there’s folks that are doing paid ads. Let’s apply that to paid ads. You know, meta is putting a lot of money behind AI. So is Amazon. Amazon is becoming a beast when it comes to the advertising world, and you can’t ignore Google, because they’ve been there for a long time. How do you think that’s going to play out when it comes to either creating better, better ads in AI versus humans that went in and probably either uploaded a bulk file of keywords and, you know, called it a day.
Julian Ghadially 11:01
Well, one of the biggest areas would be personalization. There is so much data that you can have on individuals to give them the message that resonates with them, whereas, if you’re a marketing executive or content creator, you’re giving a really broad swath method message, and then also the data that you have in general and what performs well and what doesn’t. AB testing is much wiser than what our initial assumption would have been. You know,
Rolando Rosas 11:28
I love AB testing because you find out things that sometimes you weren’t aware of. You know, we did an AB test on some thumbnails, and conventional wisdom was that one type of image would resonate more with consumers versus another. And it turns out, because of the personalization, and when we looked at the AB test, the audience that it was served up to, we found that among younger people, they preferred this type of image. And then when we looked at the demographics for older people, they preferred a different kind of image, and I think that when AI takes that into consideration, I believe meta is able to do this, and Google’s already from the paid ad, to be able to mix different assets into advertising so that it serves a personalized ad to that person, whether they’re 25 or 55 and make that advertisement more effective. I think that we’ll be entering into a whole new age of personalization, so that even the image, the voice, if you’re if it generates a voice that you like. And I’m really excited for when those two things come together, the voice aspect and then all of the data behind the text, so that you can actually hear a voice that’s to your liking on a video. Me
Julian Ghadially 12:47
too. You can apply AB testing to so many different things. I think a really cool idea in the future would be entrepreneurship, really is a little bit like AB testing. It’s like experimentation. You you test a bunch of hypotheses about a business idea, it would be pretty cool to have a AI bot that goes and, you know, tries out different messaging, sees what resonates with with people, and figures out what would be worthy for us to build.
Rolando Rosas 13:16
You know that, and that’s what we use AI to help us with the brainstorming session a lot, because that first step, sometimes it’s like, the hardest thing, you know, your mind is elsewhere. It’s over here, it’s over there. And just saying, hey, you know, I want you to help me do this task better. Or or we’re brainstorming right now, and are there some things that we’re not thinking about, and we’ve used it in house for a lot of that, and sometimes you realize there’s no way for your human brain to process all the keywords, right? Like, what would be exciting, what would be interesting, would be fascinating, would be emotional. And as AI improves over time, I can only imagine it being so human like, and this is where your expertise in this area comes into play. You know, will we get to a stage where AI will be very human, like, in its in its neural network, or its processing of data, and then spit that back out for a human? You
Julian Ghadially 14:11
know, I don’t know to what extent it’s it’s going to be like a human versus its own way of thinking. I imagine, just like we have diversity in teams, people who think differently, the AI is going to think differently, right? But I do know that there’s a lot of different dimensions of intelligence where hopefully humans will have their domain, where we do better. Otherwise, it’s a pretty dark future.
Rolando Rosas 14:37
Well, what do you let’s jump into that about the roles of humans, and right now, one of the biggest fears is that jobs would be replaced by AI. You know, there’s a lot of discussion around that, and I’ve seen different posts where engineers think, will AI kill software engineers and their jobs? What are your what’s your perspective of. Since you know, you’re deep into the llms and all the different algorithms that that go behind the scenes in putting this together. Well,
Julian Ghadially 15:07
yeah, it’s, it’s a it’s a scary world, the future of sociology, the way the Well, I think one of the biggest problems out there is if you have access to AI, and you have access to the best education, and you have access to capital to invest in these and to take time off to study, versus if you don’t, that’s a big discrepancy, and we need to really be careful in our future and when we as we build our future society, how we solve for that? But at the same time, there’s a lot of examples. Of automation coming along and not actually removing jobs, but enhancing jobs. And I think a really good example is with banks. ATMs came out, and they thought that there would be a lot less bank tellers, but it turns out that over, I think the 10 years after the ATMs came out, you ended up getting more branches, people like the convenience and people started, you know, interacting with their bank for the things that were more important. I believe there was less job loss there than than was expected. Also with with my company, we are enhancing the the role of a seller, or making sellers more efficient. It doesn’t stop them from going out and looking for arbitrage opportunities. What it does is it it gives each entrepreneur more power to find to be more profitable. So that doesn’t really take away the job necessarily.
Rolando Rosas 16:35
So it sounds like we’re going to be shifting, you know, into different roles, where maybe you were the pooper, scooper, chasing after horses, horse and buggies, and that was your job. Now your job is going to be a little bit different when it comes to AI and how you do your job. Is that the way? What I’m hearing you say, the
Julian Ghadially 16:53
way you do your job? Yes, will change 100% I think there’s a lot of repetitive tasks that we won’t have to do and things that we don’t want to do, right? And
Rolando Rosas 17:03
I’m ready to, you know what I’d love to do at some point, I’d love to just get rid of email. If there was something of a way to I have become more of a fan of instant messaging. So whether it’s teams, Microsoft Teams, or slack or Google Hangouts, the ability to instantly communicate, and maybe I don’t need an answer right now, so asymmetrically, just here’s what I need, and then when I get the response, I get it. Email is just just cluttered, just it seems clumsy. You don’t know if you’re going to get a response ever right with with instant messaging and instant communication, or unified communication, as it’s known in the industry, you get a different way of communicating, where email, it’s kind of a black box. So at some point, my hope would be there’s a way for AI to either we eliminate email altogether or, and I know a lot of email marketers are not going to like that, but or enhance it in such a way so that only the important things that I need to pay attention to are surfaced, and that’s what I actually need to read, rather than 80% or 85% of what I get in email today.
Julian Ghadially 18:17
Well, that is one thing that AI is really good at, is opportunity finding. So you take a large amount of data and you figure out, what is the thing that I need to pay the most attention to, what are, what are the opportunities for our company, for example, that would have the most impact for email? It would be, what are the emails I have I should read that will have the most impact for my job? I think that AI is really good at doing that, because you have it’s very efficient at processing that large amount of data, and when you review it as the human in the loop, you are processing a much smaller subset of data, in this case, emails. What’s
Rolando Rosas 18:52
your view when it comes to AI and humans? Is AI ready to beat the humans when it comes to different tasks? Or are we too early in this game?
Julian Ghadially 19:04
So no, already I am seeing multiple domains where AI is beating human, humans at different tasks. And I think one of the areas that is most ripe for AI to win is whenever it can generate its own data. So and you can simulate. You know, experience after experience. That is how AI learns. It’s constantly getting smarter every time. So games are a really good example of this as a little simulation ground, but it applies to the real world as well. There was a team that played the there’s a there’s a game called go. It’s very famous in the east, it’s considered very, very difficult for a computer science as a computer science problem, but the team at Google DeepMind had the AI just play millions, billions, trillions of times, and through that, and every time there’s an outcome, you win or you lose. Yeah.
Rolando Rosas 25:00
Is in sales. For those folks that are in sales, how can they apply this to enhance the organization or the sales team?
Julian Ghadially 25:06
It is an end to end, different way of thinking. Okay, from the beginning of getting your lead list to reaching out to them, personalizing your message, submitting an RFP, and then at the end of the day, after you’ve had conversations with, you know, 100 100 different sales reps having conversations with customers throughout the day, all the insights that you get right? Let’s start with the lead list. For example, AI is really good at scraping information online. There was, there was a company that I was talking to that was trying to market to some some acupuncturists. And the bot can go to every city, look at all the acupunctures, acupuncturist in each city, go to the About page and generate some really good data on what kind of acupuncture each acupuncturist practices. And then that ideal customer profile can be used to figure out these are the top acupuncturists that you need to be talking to. Then the large language model can draft a personalized outreach message, which you can just fire away, or you can have a review stage if you’d like. So the AI is talking to the unqualified opportunities, and then you spend your time on having the real conversations, doing the negotiation, etc. And with those conversations, like I said, with the insights, you’re recording everything, you’re transcribing everything. And you can see, for each different persona. What are they responding to? What are their problems, their pain points, their their pains, their gains, right? That is so transformative in terms of how you go out and sell, because now you really understand so many different customer types at a much deeper level based off of, you know, 100 times more conversations than you could have as an individual. So
Rolando Rosas 27:03
it sounds like you like AI will save not only the the task of that you’re doing, but then you can perform higher level tasks when you have these lower level tasks, maybe lead scraping, you know, identifying targets, what’s the right audience, and then the human gets involved, performs the negotiation. And
Julian Ghadially 27:24
it’s not just about efficiency, it’s about outcomes. My I have a friend who has a company called procure spark, and they generate text for RFPs, they draft the RFPs, and it allows teams to be a lot more. It’s not about efficiency, it’s about actually closing more deals, because you’re able to reach out to more deals.
Rolando Rosas 27:43
Yeah, so closing more and you know, if anybody is as in sales, that’s the bottom line. How much are you closing? What’s your closing rate? And if I can close more deals, then I’m a better salesperson. The sales organization, you can generate more revenue per person. I think that would be a win, win under that scenario for sure. You know, I can’t imagine who would not want their sales teams to close more deals, right? It’s,
Julian Ghadially 28:13
it’s all, it’s the whole value proposition. You get more, you get the outcome, and you get the efficiency.
Rolando Rosas 28:20
And so how do organizations when they’re looking at AI, you know, there’s a there’s several different ways of using it. Is this something that you know, you just, you know, put us plug in the software into your system, because you you’re talking about gathering data. So I would imagine you need conversational data. So calls being recorded, no zoom does that today. With the their call center software, they’re recording the calls, and an agent is able to intelligently identify certain things that are happening during the conversation, with some technology like that, that takes the calls, take the knowledge base of what’s happening, and then inserts it into a script or or a role play of some sort, so that the outcome is is more likely to happen that you’re looking for?
Julian Ghadially 29:05
Well, there’s a lot of barriers in implementing like we’re talking about, a lot of really, like, you know, AI is going to be the future. A is going to be great. But let’s talk first, generally, about where we are today. It’s May in 2024 there’s a lot of barriers, there’s a lot of reliability issues. I think that’s the biggest one. I’ve tried to build different large language model applications to automate different tasks. And it’s it’s great, it works like 90% of the time, but that 10% is what causes you to make a critical mistake where all of a sudden your customers are upset, or, you know, you, let’s say, in my domain, you bought the wrong product, and you had a massive, you know, multi $1,000 loss, right? So, reliability, cost, latency, the cleanliness of your data, the 10. On your team, all these things make it difficult for a business to implement this stuff. And I’m happy to also share some of the traps that I’ve fallen into too, if that’s helpful,
Rolando Rosas 30:11
sure, because I think that if you’re listening to this and you’re thinking, you know, maybe we can get a little more efficient, we can get a little more lean, but you’re saying, Hold on there, not so fast, because you need to think about this a little bit more in a planned way, because the data may not be 100% like you said, you could fall into some traps, and all of a sudden you make some bets that are the wrong bets. Well,
Julian Ghadially 30:37
I do want to encourage people to play around, right? I think the easiest way to play around with a use case is just give, give ChatGPT an example of the type of data it would be processing for, let’s say, an automation task, and see how well it does. So that was my approach. I started off with just giving different examples and seeing how well it performed. But the first trap that I fell into is you end up with this really clunky output where you didn’t design it. You just, you just started with, here’s, here’s some of the input that I think we’ll be going through it. And let’s try to, you know, engineer some things around our application to handle the input that we know about, it’s much smarter to generate. To start with, Okay, here’s all the types of input that I will be getting, and here’s a bunch of clean data on the right answer. When you have a big, clean data set like that, then you can fine tune the prompting. You can fine tune the model right? You can engineer for all the different types of bad input. That is a much more streamlined flow. It is much, you know, it’s more, much more intentional than if you just dive into it and start building so that, I think that’s one of the first traps. The other trap, I think it’s common, I for, for people to fall into is garbage in, garbage out. We all know you need clean, clean data. And
Rolando Rosas 32:09
what does that mean? Well, you’ve said it a few times now. What does clean data mean? What is that?
Julian Ghadially 32:16
So you’re using, you’re using data to train the the model. And if you’re teaching someone and you tell them something that’s false, they’re going to learn the wrong thing. If, also, if people sometimes are so high level and not oriented to the detail, if you look at the actual data you’re feeding your model, what would you be able to interpret it. I think one of the common use cases for for large language model applications is, you know, reading the internet. And if you parse HTML data into text, and you try to read it as a human, it sometimes looks there’s no Paragraph Spacing, there’s no sometimes there’s no periods within a sentence. Sometimes, like you know, if there’s a list of 10 products, those products become one big sentence, totally uninterpretable. So how do you expect the AI to learn off of that crappy, dirty data?
Rolando Rosas 33:17
Very fascinating. So is that reason why, when you get results you ask any of these chat bots today, you say, Hey, what happened in 1974 and June 1? Blah, blah, blah, blah, and it may spit out a correct answer, or it may spit something completely inaccurate or hallucinate, as the technical term is known.
Julian Ghadially 33:38
Well, there’s a lot of things driving that, but with a model that’s trained on the internet, the internet has a lot of different domains where there’s, you know, conflicting information. I mean, yahoo answers, yeah, I don’t trust I just don’t trust it, right? I would hate to have a model that is trained on forum data advise me on something that’s really important, where I’m trying to, you know, retain something and learn from it, right? So you have to be careful with what data you’re feeding in.
Rolando Rosas 34:13
Well, I think that’s a it’s a cautionary tale about using, and we’ve used AI, and sometimes it doesn’t spit out. It sounds it does the when you read it. Oh yeah, that sounds right, but it’s not right. And so it’s still, we’re not there yet, where on ChatGPT. I know they just released a new version, and we’re playing around with it. And we got some, I got some for you here in not right now, but I’m going to bring it back up. So we’re going to go back to ChatGPT in a second. But what I want to also know is, where do you think we’re going with this? Like, let’s say, 10 years from now, you looked at your crystal ball. What do you think we’re going to be with AI society in the workplace?
Julian Ghadially 34:57
Did you see the release of GPT four? Oh.
Rolando Rosas 34:59
Mm, hmm, I sure did. I saw the demo they did with the translations and and sitting around the table and asking it questions.
Julian Ghadially 35:07
I mean, you’re talking to your computer, you’re getting you’re getting real life responses, that is, it’s just moving so quickly. I it’s hard to for me to imagine what the world is going to look like in the future, but intelligence will be a commodity that we’re tapping into too. And there’s there’s just another factor. There are so many entrepreneurs out there that for every single task, are looking for ways to enhance it, make it more efficient. Those solutions are there’s nothing that is going to stop that even with the intelligence we have today, right? So then you add in the added layer of these models are just getting smarter. They’re getting more efficient. The base neural network components are getting better at getting trained on the data, right? So I think you’re going to have these agents that are playing together in the different roles that we were talking about to be a member of our team. We’re gonna have, we’re gonna have humanoid robots as well, with robotics where, you know, we can pick up heavy things, not break our backs, you know, in distribution centers, in in manufacturing centers, warehouses. So hopefully the world will be, you know, will be still the human interface that we’re used to today, and the bots will live in our world, but we will see, and there’s a lot of, there’s a lot of issues too, around ethics, around like, I’m worried about the idea of a conscious being being enslaved, right? I would be an ethical nightmare to be stuck in a box and answering boring query after boring query, right? So we have to have control and controls in place for the ethical considerations and also for the for the, you know, world ending apocalyptic scenario,
Rolando Rosas 37:04
it’s got to have a switch to turn it off and turn it on. If, if we’re concerned about that, and you talked just about something that I want to jump into right now, which is the team member aspect of it, I want, I want to play a role with you. I’m going to bring somebody on here that has been listening to this conversation. We like to call this team member, Eddie. It’s an intern. It’s that we’ve given a name here. I’m going to ask it to join our conversation. Eddie. Are you there?
AI 37:35
Given all the advancements and impacts we’ve discussed, what do you think is the next big frontier for AI that businesses should start preparing for now. Well,
Julian Ghadially 37:44
we’ve talked about a lot of different use cases, right? Automation is, is, is a huge use case. I think that that is a frontier that is more difficult to grasp. Um, one like so there’s a difference between the next frontier that will be in the future versus what we can do today. But, you know, you picture, you know, Amazon’s facilities that process things automatically with robots. So automation is a big one. Personalization is a huge one. Content Creation, decision making, right? Every single job that’s out there has some way of enhancing it. That’s the incredible piece. I think even today, though, in all those categories, there is a way for AI to play a role in automation where you can’t have reliable reliability issues, you can still have aI doing all the data processing and finding the opportunities that you know the highest opportunity set, what I like to call gold mining. So instead of automating the process, you you feed it the same data that you would as if you were automating it, and it will provide the highest, top tier recommendations for how to add value, right? But yeah, so automation, personalization and being having a role in enhancing pretty much every single task that we do
AI 39:07
that’s insightful. Julian
Rolando Rosas 39:12
and you know around here, we call this technology Eddie, which is ChatGPT, just because ChatGPT is too hard on the tongue to say ChatGPT as much as we talk about it and use it in our organization, but we thought we’d roll out, and I didn’t tell you we were going to do this, but you you like, danced around it, and that’s exactly the the demo that you were talking about, and having that very natural conversation. It’s funny that it picked up on that and went back to kind of a previous question that we asked it about the advantages and how do you use it? What do you think about Eddie now, or chatgpt when we use it like this? Well,
Julian Ghadially 39:45
it said it was listening to our conversation, but it did ask me a question that we had already talked about, right? So I think that was, you know, it’s constantly going to be getting better. I. Right? I would rather right now. I would rather talk to you. Also, there was a huge latency issue too, where you asked the question, there was this long pause. But it’s incredible that they have it out here, and this is their first version of the product, right? Like,
Rolando Rosas 40:12
yeah, it’s only gonna get better. You know, it wasn’t that long ago that this just ChatGPT itself was on the market it came out, the voice capability is improving, and I can’t wait for the day so we can get back to the earlier conversation that where the actual voice sentiment, so Right? So it’s in the background, taking the text, right, and interpreting the text, but combining the inflection of the voice. You know, if I speak in a certain tone, it means angry, right? This means sad. This means that. So combining the two to make that more effective at understanding. And that may be happening already, but I’m not, I don’t know. I haven’t seen that information yet, but when it combines the two, I think it’ll be even more powerful at bringing back an ad, like bringing back a response, like, Why do you sound so angry? All right, Eddie, I’ve got a question for you when it comes to opportunities in AI, things like gold mining and improving the sale process, sales process, what are some other aspects of AI that can help enhance sales or business processes so that AI can be used effectively in the workplace.
AI 41:23
Great question, Rolando, beyond gold mining and improving sales processes, AI can enhance business operations through predictive analytics, automating routine tasks and personalizing customer experiences. Predictive analytics can forecast trends and customer behaviors, helping businesses make proactive decisions, automating routine tasks, frees up employees to focus on more strategic activities, personalizing customer interactions can increase engagement and satisfaction by delivering tailored recommendations and support these aspects of AI can significantly boost efficiency and effectiveness in the workplace. I love
Julian Ghadially 42:02
how much it’s a proponent of its own. I agree with most of that. But today, the aspect of of automation, it’s very difficult to have a process be automated end to end, because if there’s any reliability issue, you’re on the hook for the mistake. I think a big area for AI entrepreneurs to focus on is is like automation, but not quite it’s opportunity finding. So and the best areas to to automate or to find opportunities in are the areas where you can process a large swath of data and then have a human review the top opportunities. So I like to call this gold mining, or finding a needle in the haystack. But essentially, and that’s, that’s what we do with Margin Geek, is we, we look at, you know, 10s of 1000s of product opportunities, how profitable they would be to sell online, and then the human in the loop portion is just looking at the the top 50, the top 100 opportunities, and validating, yes, that’s correct. No, that’s not. So when you go to sell that product, there’s no, there’s no error, right? The there’s only the human error, right? Um, which you know is the status quo. Anyways, I like
Rolando Rosas 43:19
that, because you know when, when it comes to number of tasks I’m thinking about like in the sales process, there’s always repeatable tasks that you do, and if the let’s say there’s 10 steps, but seven steps can be done by AI, and I’m just coming in at step, step number eight, nine and 10 to essentially close, negotiate or have the data at my fingertips that I need in order to to more rapidly chase the right opportunities. I think it’s completely valuable. You know, there are we’re drowning in data, and so understanding how to apply it, understanding how to bring it in at the right time for what I need would make me much better at what I do today. And I, like you said, AI can do that today, but it needs to have reliable data. We need to understand how we can use it in the workflow process. So I’m excited about this being plugged in and making us just a better machine for us from a sales perspective, versus boy, we got to go look through that spreadsheet, Julian, we got to go jump into that spreadsheet and look at, you know, 5000 rows and do pivot tables and find out which one has the best correlation. And, you know, and then make a decision on all that.
Julian Ghadially 44:36
I mean, we’re saving sellers hundreds of hours of the manual task of looking at 1000 different products, 2000 3000 different products, right? Even an order of magnitude greater than that. So, like I said, opportunities where there’s tons of data. The reason why the sales use case works so well is truly because you have so much information on each business that you’re. Trying to market, or each person you’re trying to market, and you you can allow it to operate on the low risk opportunities, and then focus your time on the select few high value leads, so that, yeah, that dynamic is important to to remember Julian, we love
Rolando Rosas 45:16
to do this, this little segment we call rapid fire, where we ask you questions and what’s in your brain, and whatever is there is your answer, not right or wrong, it’s your answer. So Ori, go ahead and take off the rapid fire segment for us. Awesome. All right, here we go. First question, Walmart versus Amazon. Amazon is just bigger. Bigger. Bigger is better. Amazon
Julian Ghadially 45:51
has, has really nailed so many aspects of the customer experience that Walmart is really struggling to keep up with. So, you know, people, people just, they know they’re going to get low prices. It’s going to be convenient. It’s going to arrive in two days. There’s a large flywheel of lots of sellers out there. I will say Amazon is making sellers lives very difficult, and that is not smart for their long term competitiveness, because if they lose their sellers to Walmart, they lose their their deeply entrenched competitive advantage, which is that that flywheel,
Rolando Rosas 46:25
right, the flywheel will start slowing down. We don’t want that. We don’t want that, especially being we’re on Amazon as well, and we don’t want that to happen. Favorite social media platform, oh, I’m, I’m
Julian Ghadially 46:38
a millennial. Millennials use, use Instagram. I also hate Instagram because it has turned into Tiktok. It’s turned into digital crack. And I’m a really big proponent of focus. And when you’re constantly exposing yourself, your brain to dopamine rushes. It. It creates this cycle of wanting to turn back to your phone and and have another dopamine rush. So I think my favorite social media platform is Strava. You know, you just rava is for, it’s, it’s for running and biking and swimming, hiking. It’s not a big thing. I just post my runs. I don’t really do much social aspect. Sometimes my friends will comment on it, and I’ll just, you know, have a small, small little conversation drama,
Rolando Rosas 47:29
cycling. All right, I’m gonna have to note that one for the brain bank. I like cycling, but I didn’t know there was a thing for it called strap. So I’m gonna record that for later. Favorite piece of tech.
Julian Ghadially 47:42
The technology has has been out there, you know, from from the wheel to the to fire to computers. But I have always just been fascinated by entrepreneurs who create something that makes a leap, an exponential leap in the way we do things. What I’d be most excited about is, I care about the world, the Earth, our planet, if we can make farming more and creation of food more efficient, where we don’t have to use pesticides, where we don’t have to, let’s say, where we can make vertical farming a reality. We don’t have to use land, and we can, you know, we can use power that is generated, you know, in clean methods. That is, that would be very exciting. I can’t wait to see that
Rolando Rosas 48:27
awesome clean power farming. I think there’s, we had a guest on couple months ago who suggested that a lot of those buildings where that are, a lot of them today that are empty in downtown, in some parts of America, some are very there’s very low occupancy already. Since remote work really changed all of that that we could convert some of those buildings into just that urban farming, vertical farming, housing and the rest where, you know those buildings are going to lay it away, so if they don’t put something in them. But I think, I think there’s people in different cities already starting to work on that. So let’s hope that that’s the case. You know, we get more, less less food deserts around the world. So that’d be a good thing. Empowered by clean power too. All right, first thing you reach for in the
Julian Ghadially 49:15
morning, oh, man, I reached for my phone. It’s so bad. I wish I could just, like, get out of bed instantly. I’m
Rolando Rosas 49:27
not a morning person myself, so I do reach for the phone as well. You it’s,
Julian Ghadially 49:32
it’s very, you know, I hear all these people talk about, you know, if you want to start your day off right, you have to, you have to have a routine in the morning. I I wish I was that good at having a discipline right when I wake up, but I just my brain needs a moment to wake up.
Rolando Rosas 49:48
I know I need a hot moment. I have blackout shade so it’s completely black, and it takes me a while. Started using one of these little wake up lights to make it just a little bit easier for me in the morning, because I. Evans, no, I’m not. I’m not the best morning person. I’m but trying to get better at it. I want to know, and this is something I always find intriguing, because every guest brings something new to the table when it comes to this next one, Game Changing book, a
Julian Ghadially 50:17
lot of books have been impactful to me. I love the the book abundance, because for my life and figuring out what problems I want to work on, it talks about the types of problems that really matter, things where, at least from a business perspective, since I’m in business, if you are able to make something more efficient. Everyone adopts it that kind of, those kinds of game changing opportunities, like around health care, around food and agriculture. But then also, let’s see emotional intelligence, by Daniel Goldman, is a great one. What else? Oh, this is this one. Is this one is core to my philosophy in life. But I picked up a book on humanism one day, which, after there’s in America, a lot of people pronounce religion at a certain point, and you need something to fill the place, if, if, if that’s what you’re doing. And humanism is all about, how do you build a life framework for yourself to be happy? What are your values? What? What do you care about? Why are you here? What’s your purpose? Right? Really, having that be defined so that you can make your other decisions in your life accordingly, your job should be aligned with what you think is important in life, right? So, yeah, picking up a book on humanism, Buddhism also has a similar, similar themes,
Rolando Rosas 51:53
yeah, all right, we’ve got several for the brain bank on this one last one, a person you admire.
Julian Ghadially 52:00
I have always been a sucker for entrepreneurs that are able to solve something and bring something into the hands of millions and billions of people, which is why I became an entrepreneur. So in the early days, this was I was crazy about Steve Jobs. I was crazy about the Wright brothers with aviation. Elon Musk is really controversial nowadays, but I was always a musk fan because the idea of clean power, clean energy, the idea of taking humanity to another planet, and the actions that go behind that, not just his actions as his whole team. So it’s really inspiring how much impact you can have when you focus on things that have that exponential type of impact.
Rolando Rosas 52:48
All right, wonderful. Now, if somebody wants to get in touch with anything, you know what? Let me figure this AI thing, is it right for my business? Is it right for us to be using? How should somebody get a hold of you. You
Julian Ghadially 53:01
can reach out to me on LinkedIn. I think they’ll be so my name is Julian. Julian Ghadially. You can find me that way. I think my name is posted here somewhere.
Rolando Rosas 53:12
Yeah. All right. So LinkedIn, is there your preferred route? So Julian, if people are interested in what you’re doing, how can they find out more about your company? Our company
Julian Ghadially 53:24
is called Margin Geek. Our website is margingeek.com we are all about helping e-commerce sellers find the cream of the crop best opportunities to sell on. If you are scanning using normal tools or scanning manually, it is a lot of time. It is a lot of manual work. But most importantly, you’re missing out on the cream of the crop opportunities. And we’re all about scanning as deeply as as not just humanly possible, but beyond humanly possible. So our AI process is finding 25% more profit, 50% more opportunities. So that means, in an arbitrage sense, you are picking the all the cream of the crop opportunities. And I love to get to have a conversation with you. You can try it out just from the website, or you can book a little demo, and I’d be happy to show it to you. Awesome,
Rolando Rosas 54:21
wonderful. Margingeek.com, right? That’s correct, awesome. Well, hopefully you go check that out and see if that’s going to be right for you. But it sounds like if you’re spending your hat, you’re hanging your hat in terms of selling products, it’s probably something that can save a ton of time, ie, cash, as well as make you more profitable. So Margin Geek, we have it here on the screen, so you can go check that out as well. So I want to thank Julian for coming hanging out with me, for today, being such a great sport, letting our AI assistant intern Eddie, take a crack at being a podcast host and asking some questions. So Julian, I want to thank you for. That. And if you’ve found this podcast to be valuable, you want to go check out that episode that we did with Neil Patel, who talked about SEO and what it means to a business and growing it and debunking some of those common myths. And I want to invite you if you want to support this channel so that we can bring more wonderful guests like Neil Patel as well as Julian to the podcast. Go check out the links in our description where we have additional information on business services for those businesses that are looking for a technology roadmap. So go check that out. You’d help the channel a lot, and we can bring you more guests like this in the future. So I want to thank you for joining me today, and I will see you in that episode with Neil Patel.
Recent Comments