Luke:
Hello and welcome to the next session in CareerCon 23, where we're talking all about AI and how it's changing the way we learn and work. In this session, I welcome the co-founder and CEO of Sociate, Yasmin Topia. Welcome, Yasmin. Where you're gonna join us and talk about how AI's emotional edge is shaping the way we work, where we live and the way we shop. And we spend a lot of time shopping. So I'm looking forward to hearing all about this. Yasmin, on your LinkedIn, you talk about deep tech and websites. dying and changing. So I'm really looking forward to hearing what you've got to say. Quickly, who am I? I'm the CEO of hundo and I'm the host and that's all I need to say about me because this is all about the insights that you're here to share, Yasmin. So thank you for spending the time with us today and over to you to tell us about yourself, about Sociate and what you're doing.
Yasmin:
Thanks Luke for having me. It's a pleasure to be here. So a quick background then. So I'm Yasmin Topia, CEO of Sociate. I'm also a founding partner of a deep tech venture builder called Post Urban Ventures that we set up in 2016 to de-risk AI startups. So I work with a community of world-leading academics and scientists who are ahead of academia and these people are doing amazing things in the field of AI research. but like academics, they're not necessarily thinking about real world problems or those commercial opportunities. And that's where I come in. So my background is pretty solid in sales, business development. I was with Adecco and PWC where I was responsible for delivering over 40 million pounds worth of new business. And over the course of the years, I have helped build four other AI companies that are all now backed by investors like Microsoft, Jaguar Land Rover, and Silicon Valley VCs. Now, In January 2020, my chief scientist came to me with a fascinating bit of research that he'd been working on for over 15 years called curious AIs. So these AIs are curious. They know what they know. They know what they don't know. But importantly, they know what they need to find out to help humans. And what Luke had done is while he had associated emotions with the boundaries of the AIs, knowledge, just like us as humans when we're having a discussion and Luke, when you'll ask me a question and I don't quite know the answer to that, you will hear in my voice how I signaled slight shades of humility, uncertainty, even at a subconscious level, which encourages you to share more information, more data with me to help me over that bump in that conversation so I can give you an answer. Conversely, if we're talking about something that I'm really passionate about, I'll get excited. I'll get happy. Again, you'll feed off that excitement. You'll want to add to that excitement. So you give me more information to learn from, to encourage that. And this is so important when it comes to large language models and AIs that we speak to now in this world of conversational AI, because At the moment, when AIs get things wrong, when they break, we get very frustrated and we want to end the conversation. Stupid bot, you're driving me mad. I don't want to talk to you anymore. And that is literally the worst thing that can happen for an AI human relationship that's stopping interaction. So that's the core AI that we've been working on. And where I have applied that. is to transform how we shop online. So we've taken that curious AI. We've applied it to AIs that can see and speak to act as shopping assistance so that in the future, you never have to spend hours clicking and scrolling. You are simply able to communicate exactly what it is that you want. No matter how niche, no matter how strange, if you want a dress that looks like a termite mount, just ask the AI, it will find you a dress that looks like an ant.
Luke:
Sounds absolutely fascinating and very helpful. You know, we've been shopping online for maybe, come on, 20 years or something now. And actually the experience hasn't really changed that much. This sounds like a real step change in the way that we research products that are right for us and then ultimately find the one that we want to purchase. Interestingly, Yasmin, of course, people who may be watching this will have been hearing lots about AI. In the last six months and they'll think it's something new. But one thing that you've informed us of there is that it's not, it's been around for a long time and people are working behind the scenes or below the headlines and there are a lot of headlines now. So, you know, how long has it been around? How long, how long have you got, been in AI? How did you get into AI?
Yasmin:
So I've been in AI 2016 and I do not have a technical bone in my body. And this is what's really important around using technology and advancing technology, because there is a level here where there is decades of work when it comes to researching and getting the technology to where it is. And you're absolutely right. AI is not new, but how it has been working, so far it has been really cumbersome. So the step change that you are talking about and the headlines around generative AI is as a result of the fact that previously AI applications to work would need so much human input, so much tagging, the experience I'm talking about as well, online, that has not been possible because at the moment, if you go to a website and you say, I need a dress for a picnic. If nothing's being tagged with the words picnic, then the AI can't show it to you because it's just doing that matching of the words. So now the intelligence is that the AI knows what a picnic is. It knows what you should wear to a picnic. It can look. So in our case, our AIs have vision as well as having the ability to... to interrogate the language data. And so it can say, well, is this within scope of a picnic and then pull it out. But how did I get into AI? So I was invited specifically to join this group of academics and scientists specifically because I built up my career in... in the real world, problem solving. So you don't necessarily need a technical background to work in AI. Having real world experience, being able to problem solve is as important as being technical because it's only when you marry the two things together that you create solutions for the world that are actually useful.
Luke:
Yeah, that's a really interesting and very helpful guide for people who are watching this, because they may be thinking that in order to work in the AI sector or in businesses that are AI-led, that they need to have a technical background. But actually, you're saying, no, that's not the case. If you approach this with critical thinking and a problem-solving mindset, there's lots of opportunities for you.
Yasmin:
Yeah, absolutely. Just taking the example of the curious AI, there are so many applications of how you can fit it in. And what I was able to do is go, well, actually, this is a big problem. People are going to pay for this problem to solve, to be solved. It can potentially change your world. Let's go and tackle this problem. Somebody else would come along and with their background, with their interests, be able to identify something else that they can solve. with the same sort of technology. And I think that's the way of looking at it, that not necessarily thinking about AI as AI, but as a tool, as capabilities. How can you now take this groundbreaking innovation? Where can you apply it? What are you also really passionate about? Because that's where the authenticity and the ability to solve the problem really comes. And the fact that then you can stick with that problem and you can get the results that you are aiming for.
Luke:
I want to ask you how you came across the problem. What was the problem that you identified? How did you come across that problem? And you started to answer it there. You're a shopper. So how did you come across the problem and start to think about solving this?
Yasmin:
Okay, so back in 1994, when I was 10, 11, I watched Clueless. And I don't know if your audience have seen Clueless, but Cher Horowitz, who is a hero of mine, has a digital wardrobe, where it is basically got contents of a wardrobe and it makes styling recommendations. Now, at that point, nothing like that existed. And I waited for years. for something like that to come into being. By 2020, I was thinking, how can we have such smart people? How can we have this AI with the ability to do what it can and nobody has made shopping and styling and dressing as easy as it is for Cher in Clueless? And so that's sort of the problem element kind of in my head came from. But... The hard problem that we are solving is if you go onto something like Farfetch, for example, and you search for a red dress, you get 1,073 results. You probably get about 25 pages. Now, what we know from research is that 79% of customers, including me, will have dropped off by page three because, oh my God, I am bored. I don't want to go through all of this. And so if, as a customer, find what you are looking for, you simply can't buy it. And that's a crying shame for retailers because that dress is in there. If only I got to it. And so the thinking was, how can we surface for customers the thing that I actually want to look at? The most natural thing to do is just to be able to convey what it is that you like. And the best way for brands and retailers to learn about what it is that you want isn't necessarily by stalking you or collecting data and cookies. Just ask, ask your customers, let them tell you, and then give them what it is that you want. So that's really the philosophy of what it is that I'm creating. And this is what I've always kind of wanted to replicate. And we are just so lucky that we are now in a world and I've got these smart people behind me that can help me recreate that experience.
Luke:
I love it. I love it. So you're really using technology to solve the problem that a consumer has, which is finding that right piece of that right item for that right occasion. And the problem that the retailers have, which is knowing that they have the right items for that occasion, but not being able to get through the tech layer to communicate and give it to the customer. So I really see the problem. And it's amazing to see someone solving that. Now, as you've been developing the platform, can you share some fascinating insights? that you've gained from harnessing the voice tone and the emotion from the AI technology that you're developing.
Yasmin:
Sure. So what we've released so far is in text, but what we have done is lots of experiments around emotional AI and the impact of voice tone on user engagement. And what's really fascinating is that by giving the AI in A-B trials, the ability to adjust its voice based on its confidence, even to the human ear. In really kind of imperceptible ways, the unconscious impact, we found that when talking to the curious AI, the customers were 23% less frustrated, they were 21% more engaged. And what that led to were longer conversations. And when you have longer conversations, you are sharing more data, which allows the AI to learn faster. So without you as a human knowing it, you're teaching the AI and it's learning faster. And that is a complete world away from, as we kind of touched upon already, AI tagging. You know, you've got these massive centers and people sitting at their screens right this minute, tagging pictures, tagging documents to make it digestible by AI. Whereas our approach is to create the world's most engaging AI, so it can be the fastest learning AI. So that was a very happy finding because it proved our theory that if the AIs are able to express emotion around the boundaries of their knowledge, then humans will engage for longer, they will share more. And so those AIs become smarter and smarter and smarter naturally. But, um, something that we weren't prepared for, and I don't know if there is a sociological commentary around this, is that we found that women had a 30 point more positive reaction to the AI showing that emotion compared to those who identified as men. So even though men ungraciously did engage for longer and did provide more information, which obviously helped the AI, when we took the qualitative feedback, men felt that the AI was the emotional AI or the curious AI was less intelligent than the non-curious, the AI that just delivered everything in a certain tone. And didn't consciously like that so much. So that's really, really fascinating around, as I said, a wider social commentary around how different groups engage with curiosity and the willingness to show humility or in a slight dip in confidence. And again, it was something that then we had to take into consideration around what problem we solve, or who we target and fashion shopping, apparel, that market is predominantly focused towards people who identify as female. So this is where early research in product development plays a huge part and is really, really important because you won't always come with an obvious answer. This is a technology. This is where you solve it. When it comes to deep tech, you almost have to reverse that. You have to think, well, actually this is the technology. Where does it go in the world to make the greatest impact?
Luke:
Since in the time that you've been developing it, how have you seen your users adapt? Because we've seen AI really storm into our consciousness in the last six to nine months in the form of generative AI. And we're all amazed and hyped about it for three months, four months, five months. But we've adapted very quickly. We're using it in our day-to-day working lives now. So there's two part question is How are you seeing users adapt to it? Where initially they might say, well, that's a bit too much information. That's a bit too personal. Are they quickly adapting to it? And then subsequent to that, within Sociate, how have you adapted as a team to using AI in your day-to-day work?
Yasmin:
Okay. So in terms of how quickly customers adapt, I think you're completely right. The way that AI is packaged up now, for example, chat GPT, these conversational experiences, they're so natural that, and because they're so natural, it's instinctive. In retail, however, it's still very early days. So people don't trust. the search bar and people also have a little bit of an ickiness around the idea that there's an AI maybe following them around, you know, and that's just natural. Again, we've come to the conclusion that actually AI needs to be invisible. So how do we put it behind the search bar? So to you, everything looks the way that it should, but suddenly everything works a lot better. So that's the path that we've taken that make it imperceptible in terms of its presence but just make everything a hundred times better in terms of performance. And then you can start using subtle tricks and prompts to help humans just know what the technology is capable of. So I think it's very important to bear in mind that everybody is on a journey. And to make those steps gradually and help through these little tricks and tools and prompts to help people understand what that technology is capable of, and so that they can build confidence as well. Internally, how are we using AI? I mean, chat GPT is great, for email drafting, it's great for proposal drafting, you know, the ability to be able to do a mind dump. What is it that I want to say? I don't have to think about how I'm going to word it. It's just about, you know, the substance of what I want to get across. And then, chat GPT can just simply take that and turn it into a nice email. You know, I need to send an email to an advisor. This is what I want to say to him. I want the tone to be professional. and friendly, the AI will turn it into that email. It's taken me five minutes to draft an email that would have taken me 20 minutes. Now that's a huge leap in productivity. And so personally, this is what's been great about these tools.
Luke:
It goes back to that point of problem solving. AI isn't going to do all of the job for you, but it will help you to ideate solutions to the problems that you identify, whether that problem is how do I tone this email or, you know, please summarize this long article because the problem is I haven't got time to read it or please summarize it for me. It's helping people to ideate solutions of all sorts of different types, which is fascinating. I want to switch a little bit away from AI and actually into shopping and e-commerce, because there'll be people who are watching this that are more interested in shopping and e-commerce. And I know that marketing courses in university and college talk a lot about certain subjects like branding and messaging, but not necessarily about performance and ROI, you know, a lot about that. I know a little bit about that. You know, we must get a hundred users to the site. A number of those will purchase for an average order value, which will give an ROI. Can you talk a little bit more about that so that people who are studying marketing and want to go into fashion can prepare more for the reality of driving performance and what that means?
Yasmin:
Yeah. Um, ROI is everything. You know, it's, it's the reason that a company will choose to work with your technology, if you can't somehow add to the bottom line or make significant savings, then essentially you don't have a business and the ROI that we are targeting is really around generating revenues through really deep personalization. So by making websites conversational, you personalize the ability of websites to serve people around that specific need in that specific moment. I need this for that. And if you've got millions of customers, then there is potentially, you know, hundreds of thousands of different needs. And allowing people to communicate that means that you can serve your customers in that personal way. And you can also learn as well by allowing people to just talk about what it is that they want. That business intelligence to help retailers, brands then around supply, demand, how much they should manufacture, when they should reorder, all of that is so, so important. But serving customers. So just to give you an example around the performance of our AI, at the moment, our client has a very famous search engine that they use for retail, a search and discovery tool. And that search to purchase conversion sits at 0.7%. By introducing our AI. That search to purchase conversion went to 1.4%. The stock is the same.
Luke:
So for every 100 people that visit the website, 1.4 people actually purchase something. Or for every 1000, 14 whole people purchase something, which is a huge difference from the original number, double.
Yasmin:
Exactly, it's essentially double. And that's crazy when you think about the fact that stock is exactly the same, the brand is exactly the same, everything else is absolutely the same. But suddenly, it's double the income. And that's the sort of thing that marketers and businesses want to see. So, you know, personalization is everything and the way that we personalize is, yes, we allow people to convey what it is that they want, but also with AI, you can start learning through purchase history.
Luke:
Amazing. So what you're talking about is AI is personalized and has memory. It has memory so that it can give you predictive, it can predict what you want and make predictive suggestions to you. And ultimately with all of that, you're optimizing for time, aren't you? You're optimizing for my time as a user who is looking for a specific thing, for a need that I have.
Yasmin:
Exactly, exactly. You know, I really do imagine this world where I can just pick up my phone and say, I've got an event to go to in a couple of days. It's a networking event. It's casual. It's formal. I don't have anything to wear. And the AI going, how about this? And I know you're going to love this. Or maybe the AI can even look into my existing purchases and say, well, I know you own this, and this, so how about you pair it with this? And that's a whole new app. I am no longer spending all of the time clicking, scrolling, searching around. I'm optimizing and utilizing what I have. I'm buying those hero pieces to uplift my wardrobe, make it current for a new season. There's all these things that you're unlocking now with the AI's capability.
Luke:
Yasmin, we're coming towards the end and everything you've shared has been super insight into AI, into shopping, into other areas. People are listening here and they're thinking, okay, what valuable lessons and insights does Yasmin and Sociate have that I could perhaps apply into other industries, number one. And number two, what skills are really important for the next five, 10 years of a career in this space?
Yasmin:
When we started Sociate and we said we were going to create a curious AI that could see and speak, people said, don't be silly, you know, that's impossible. You can't, you can't do that. What I want to say is that actually with the advances in AI, with what's happening in academia, that barrier of you can't physically because the tools aren't there, it's gone or it's going. And the only way you can advance technology is by really pushing and looking into the realms of possibility and science fiction. So if you have ideas, don't listen to people that say that's not possible. Find out for yourself. I would really recommend, especially if you are on your way to university, to get involved in societies where you can get exposure to the people that are also in research, in deep tech, in academia, and talk to them. And sometimes it might just sound incredibly academic, but try to get to... the crux of what it is that they are working on. And I would say work backwards. So, you know, the general wisdom is find a problem and then the solution. But now with where technology is, imagine a solution and think about where the world might be in four or five years. Imagine and build for that because everything we know and how it works is gonna change. And none of us really have a grasp of what the world is going to look like in five or ten years, really, but we do know it's going to change. So just let yourself imagine and build for that world.
Luke:
I love that. So what you're saying is actually, forget the skills, be curious, have a big imagination and go for it. Be bold.
Yasmin:
Curiosity is the most valuable tool that we have as humans. It's why we decided to give our AI curiosity. Because if you've got curiosity, you can learn, you can uncover, you go down these paths. And none of those paths, even if they do come to a dead end, they're not wasted journeys because you've learnt so much and then you just kind of, yeah, branch off a little bit. So, yeah, you know, imagine, imagine what the world is going to be and let that run wild and it will all be in the realms of possibility.
Luke:
I love it. I love it. And I completely agree with you around curiosity. Be curious about the problem, why it exists, unpack it, and then be curious about potential solutions to it and why they haven't been done yet. Maybe there's an opportunity. Yasmin, in the moment that we've got left and thank you so much for this really fascinating conversation, let's end by sharing with us the one thing that you're really excited about with the future of AI in the next, you name the time frame.
Yasmin:
So to me, the exciting thing is that AI really does a lot of potential in humans. And I think that's how you see it. So, whatever your passion is, if you have ever felt that you don't necessarily have the technical skills, that goes away. And to me, it's that democratization of ability. It's, you know, reducing the barriers of access for so long, especially when it comes to entrepreneurship, you need to have the right networks, you need to have the right friends, you need to have, you know, know how to do marketing, all of that. Now, so much of it is actually within your control and you can execute. And I think that is the most exciting thing for every single one of us.
Luke:
Fantastic, I completely agree. And I know people are gonna want to connect with you, Yasmin, and read more of your content and see more of your videos where you're sharing your knowledge and your insights from Sochiate. Where can people continue to engage with you after this session?
Yasmin:
So find me on LinkedIn. It's Yasmin Topia. If you're not on LinkedIn, I am on Instagram @Sociate AI, or you can also find me on Twitter. I'm not very active on Twitter, but on Instagram or on LinkedIn would be the best place.
Luke:
And on your website.
Yasmin:
And on my website, yep, you can kind of click that button for a demo and at the moment that email comes straight to me.
Luke:
Very good. Thank you so much, Yasmin. That's been a really fascinating session. You've shared so much insight, not only about AI, but about shopping and about the skills needed to come into this space. I'm sure people are gonna wanna connect with you. So thanks for sharing your details there. And of course, thank you very much and good luck with the future of Sociate and saving us all time, finding the relevant things that are gonna help us look amazing or solve problems that we have on our daily lives.
Yasmin:
No thank you very much for inviting me. It's been a pleasure.
Luke:
Thank you so much. Take care.
Yasmin:
Bye.
Luke:
Hello and welcome to the next session in CareerCon 23, where we're talking all about AI and how it's changing the way we learn and work. In this session, I welcome the co-founder and CEO of Sociate, Yasmin Topia. Welcome, Yasmin. Where you're gonna join us and talk about how AI's emotional edge is shaping the way we work, where we live and the way we shop. And we spend a lot of time shopping. So I'm looking forward to hearing all about this. Yasmin, on your LinkedIn, you talk about deep tech and websites. dying and changing. So I'm really looking forward to hearing what you've got to say. Quickly, who am I? I'm the CEO of hundo and I'm the host and that's all I need to say about me because this is all about the insights that you're here to share, Yasmin. So thank you for spending the time with us today and over to you to tell us about yourself, about Sociate and what you're doing.
Yasmin:
Thanks Luke for having me. It's a pleasure to be here. So a quick background then. So I'm Yasmin Topia, CEO of Sociate. I'm also a founding partner of a deep tech venture builder called Post Urban Ventures that we set up in 2016 to de-risk AI startups. So I work with a community of world-leading academics and scientists who are ahead of academia and these people are doing amazing things in the field of AI research. but like academics, they're not necessarily thinking about real world problems or those commercial opportunities. And that's where I come in. So my background is pretty solid in sales, business development. I was with Adecco and PWC where I was responsible for delivering over 40 million pounds worth of new business. And over the course of the years, I have helped build four other AI companies that are all now backed by investors like Microsoft, Jaguar Land Rover, and Silicon Valley VCs. Now, In January 2020, my chief scientist came to me with a fascinating bit of research that he'd been working on for over 15 years called curious AIs. So these AIs are curious. They know what they know. They know what they don't know. But importantly, they know what they need to find out to help humans. And what Luke had done is while he had associated emotions with the boundaries of the AIs, knowledge, just like us as humans when we're having a discussion and Luke, when you'll ask me a question and I don't quite know the answer to that, you will hear in my voice how I signaled slight shades of humility, uncertainty, even at a subconscious level, which encourages you to share more information, more data with me to help me over that bump in that conversation so I can give you an answer. Conversely, if we're talking about something that I'm really passionate about, I'll get excited. I'll get happy. Again, you'll feed off that excitement. You'll want to add to that excitement. So you give me more information to learn from, to encourage that. And this is so important when it comes to large language models and AIs that we speak to now in this world of conversational AI, because At the moment, when AIs get things wrong, when they break, we get very frustrated and we want to end the conversation. Stupid bot, you're driving me mad. I don't want to talk to you anymore. And that is literally the worst thing that can happen for an AI human relationship that's stopping interaction. So that's the core AI that we've been working on. And where I have applied that. is to transform how we shop online. So we've taken that curious AI. We've applied it to AIs that can see and speak to act as shopping assistance so that in the future, you never have to spend hours clicking and scrolling. You are simply able to communicate exactly what it is that you want. No matter how niche, no matter how strange, if you want a dress that looks like a termite mount, just ask the AI, it will find you a dress that looks like an ant.
Luke:
Sounds absolutely fascinating and very helpful. You know, we've been shopping online for maybe, come on, 20 years or something now. And actually the experience hasn't really changed that much. This sounds like a real step change in the way that we research products that are right for us and then ultimately find the one that we want to purchase. Interestingly, Yasmin, of course, people who may be watching this will have been hearing lots about AI. In the last six months and they'll think it's something new. But one thing that you've informed us of there is that it's not, it's been around for a long time and people are working behind the scenes or below the headlines and there are a lot of headlines now. So, you know, how long has it been around? How long, how long have you got, been in AI? How did you get into AI?
Yasmin:
So I've been in AI 2016 and I do not have a technical bone in my body. And this is what's really important around using technology and advancing technology, because there is a level here where there is decades of work when it comes to researching and getting the technology to where it is. And you're absolutely right. AI is not new, but how it has been working, so far it has been really cumbersome. So the step change that you are talking about and the headlines around generative AI is as a result of the fact that previously AI applications to work would need so much human input, so much tagging, the experience I'm talking about as well, online, that has not been possible because at the moment, if you go to a website and you say, I need a dress for a picnic. If nothing's being tagged with the words picnic, then the AI can't show it to you because it's just doing that matching of the words. So now the intelligence is that the AI knows what a picnic is. It knows what you should wear to a picnic. It can look. So in our case, our AIs have vision as well as having the ability to... to interrogate the language data. And so it can say, well, is this within scope of a picnic and then pull it out. But how did I get into AI? So I was invited specifically to join this group of academics and scientists specifically because I built up my career in... in the real world, problem solving. So you don't necessarily need a technical background to work in AI. Having real world experience, being able to problem solve is as important as being technical because it's only when you marry the two things together that you create solutions for the world that are actually useful.
Luke:
Yeah, that's a really interesting and very helpful guide for people who are watching this, because they may be thinking that in order to work in the AI sector or in businesses that are AI-led, that they need to have a technical background. But actually, you're saying, no, that's not the case. If you approach this with critical thinking and a problem-solving mindset, there's lots of opportunities for you.
Yasmin:
Yeah, absolutely. Just taking the example of the curious AI, there are so many applications of how you can fit it in. And what I was able to do is go, well, actually, this is a big problem. People are going to pay for this problem to solve, to be solved. It can potentially change your world. Let's go and tackle this problem. Somebody else would come along and with their background, with their interests, be able to identify something else that they can solve. with the same sort of technology. And I think that's the way of looking at it, that not necessarily thinking about AI as AI, but as a tool, as capabilities. How can you now take this groundbreaking innovation? Where can you apply it? What are you also really passionate about? Because that's where the authenticity and the ability to solve the problem really comes. And the fact that then you can stick with that problem and you can get the results that you are aiming for.
Luke:
I want to ask you how you came across the problem. What was the problem that you identified? How did you come across that problem? And you started to answer it there. You're a shopper. So how did you come across the problem and start to think about solving this?
Yasmin:
Okay, so back in 1994, when I was 10, 11, I watched Clueless. And I don't know if your audience have seen Clueless, but Cher Horowitz, who is a hero of mine, has a digital wardrobe, where it is basically got contents of a wardrobe and it makes styling recommendations. Now, at that point, nothing like that existed. And I waited for years. for something like that to come into being. By 2020, I was thinking, how can we have such smart people? How can we have this AI with the ability to do what it can and nobody has made shopping and styling and dressing as easy as it is for Cher in Clueless? And so that's sort of the problem element kind of in my head came from. But... The hard problem that we are solving is if you go onto something like Farfetch, for example, and you search for a red dress, you get 1,073 results. You probably get about 25 pages. Now, what we know from research is that 79% of customers, including me, will have dropped off by page three because, oh my God, I am bored. I don't want to go through all of this. And so if, as a customer, find what you are looking for, you simply can't buy it. And that's a crying shame for retailers because that dress is in there. If only I got to it. And so the thinking was, how can we surface for customers the thing that I actually want to look at? The most natural thing to do is just to be able to convey what it is that you like. And the best way for brands and retailers to learn about what it is that you want isn't necessarily by stalking you or collecting data and cookies. Just ask, ask your customers, let them tell you, and then give them what it is that you want. So that's really the philosophy of what it is that I'm creating. And this is what I've always kind of wanted to replicate. And we are just so lucky that we are now in a world and I've got these smart people behind me that can help me recreate that experience.
Luke:
I love it. I love it. So you're really using technology to solve the problem that a consumer has, which is finding that right piece of that right item for that right occasion. And the problem that the retailers have, which is knowing that they have the right items for that occasion, but not being able to get through the tech layer to communicate and give it to the customer. So I really see the problem. And it's amazing to see someone solving that. Now, as you've been developing the platform, can you share some fascinating insights? that you've gained from harnessing the voice tone and the emotion from the AI technology that you're developing.
Yasmin:
Sure. So what we've released so far is in text, but what we have done is lots of experiments around emotional AI and the impact of voice tone on user engagement. And what's really fascinating is that by giving the AI in A-B trials, the ability to adjust its voice based on its confidence, even to the human ear. In really kind of imperceptible ways, the unconscious impact, we found that when talking to the curious AI, the customers were 23% less frustrated, they were 21% more engaged. And what that led to were longer conversations. And when you have longer conversations, you are sharing more data, which allows the AI to learn faster. So without you as a human knowing it, you're teaching the AI and it's learning faster. And that is a complete world away from, as we kind of touched upon already, AI tagging. You know, you've got these massive centers and people sitting at their screens right this minute, tagging pictures, tagging documents to make it digestible by AI. Whereas our approach is to create the world's most engaging AI, so it can be the fastest learning AI. So that was a very happy finding because it proved our theory that if the AIs are able to express emotion around the boundaries of their knowledge, then humans will engage for longer, they will share more. And so those AIs become smarter and smarter and smarter naturally. But, um, something that we weren't prepared for, and I don't know if there is a sociological commentary around this, is that we found that women had a 30 point more positive reaction to the AI showing that emotion compared to those who identified as men. So even though men ungraciously did engage for longer and did provide more information, which obviously helped the AI, when we took the qualitative feedback, men felt that the AI was the emotional AI or the curious AI was less intelligent than the non-curious, the AI that just delivered everything in a certain tone. And didn't consciously like that so much. So that's really, really fascinating around, as I said, a wider social commentary around how different groups engage with curiosity and the willingness to show humility or in a slight dip in confidence. And again, it was something that then we had to take into consideration around what problem we solve, or who we target and fashion shopping, apparel, that market is predominantly focused towards people who identify as female. So this is where early research in product development plays a huge part and is really, really important because you won't always come with an obvious answer. This is a technology. This is where you solve it. When it comes to deep tech, you almost have to reverse that. You have to think, well, actually this is the technology. Where does it go in the world to make the greatest impact?
Luke:
Since in the time that you've been developing it, how have you seen your users adapt? Because we've seen AI really storm into our consciousness in the last six to nine months in the form of generative AI. And we're all amazed and hyped about it for three months, four months, five months. But we've adapted very quickly. We're using it in our day-to-day working lives now. So there's two part question is How are you seeing users adapt to it? Where initially they might say, well, that's a bit too much information. That's a bit too personal. Are they quickly adapting to it? And then subsequent to that, within Sociate, how have you adapted as a team to using AI in your day-to-day work?
Yasmin:
Okay. So in terms of how quickly customers adapt, I think you're completely right. The way that AI is packaged up now, for example, chat GPT, these conversational experiences, they're so natural that, and because they're so natural, it's instinctive. In retail, however, it's still very early days. So people don't trust. the search bar and people also have a little bit of an ickiness around the idea that there's an AI maybe following them around, you know, and that's just natural. Again, we've come to the conclusion that actually AI needs to be invisible. So how do we put it behind the search bar? So to you, everything looks the way that it should, but suddenly everything works a lot better. So that's the path that we've taken that make it imperceptible in terms of its presence but just make everything a hundred times better in terms of performance. And then you can start using subtle tricks and prompts to help humans just know what the technology is capable of. So I think it's very important to bear in mind that everybody is on a journey. And to make those steps gradually and help through these little tricks and tools and prompts to help people understand what that technology is capable of, and so that they can build confidence as well. Internally, how are we using AI? I mean, chat GPT is great, for email drafting, it's great for proposal drafting, you know, the ability to be able to do a mind dump. What is it that I want to say? I don't have to think about how I'm going to word it. It's just about, you know, the substance of what I want to get across. And then, chat GPT can just simply take that and turn it into a nice email. You know, I need to send an email to an advisor. This is what I want to say to him. I want the tone to be professional. and friendly, the AI will turn it into that email. It's taken me five minutes to draft an email that would have taken me 20 minutes. Now that's a huge leap in productivity. And so personally, this is what's been great about these tools.
Luke:
It goes back to that point of problem solving. AI isn't going to do all of the job for you, but it will help you to ideate solutions to the problems that you identify, whether that problem is how do I tone this email or, you know, please summarize this long article because the problem is I haven't got time to read it or please summarize it for me. It's helping people to ideate solutions of all sorts of different types, which is fascinating. I want to switch a little bit away from AI and actually into shopping and e-commerce, because there'll be people who are watching this that are more interested in shopping and e-commerce. And I know that marketing courses in university and college talk a lot about certain subjects like branding and messaging, but not necessarily about performance and ROI, you know, a lot about that. I know a little bit about that. You know, we must get a hundred users to the site. A number of those will purchase for an average order value, which will give an ROI. Can you talk a little bit more about that so that people who are studying marketing and want to go into fashion can prepare more for the reality of driving performance and what that means?
Yasmin:
Yeah. Um, ROI is everything. You know, it's, it's the reason that a company will choose to work with your technology, if you can't somehow add to the bottom line or make significant savings, then essentially you don't have a business and the ROI that we are targeting is really around generating revenues through really deep personalization. So by making websites conversational, you personalize the ability of websites to serve people around that specific need in that specific moment. I need this for that. And if you've got millions of customers, then there is potentially, you know, hundreds of thousands of different needs. And allowing people to communicate that means that you can serve your customers in that personal way. And you can also learn as well by allowing people to just talk about what it is that they want. That business intelligence to help retailers, brands then around supply, demand, how much they should manufacture, when they should reorder, all of that is so, so important. But serving customers. So just to give you an example around the performance of our AI, at the moment, our client has a very famous search engine that they use for retail, a search and discovery tool. And that search to purchase conversion sits at 0.7%. By introducing our AI. That search to purchase conversion went to 1.4%. The stock is the same.
Luke:
So for every 100 people that visit the website, 1.4 people actually purchase something. Or for every 1000, 14 whole people purchase something, which is a huge difference from the original number, double.
Yasmin:
Exactly, it's essentially double. And that's crazy when you think about the fact that stock is exactly the same, the brand is exactly the same, everything else is absolutely the same. But suddenly, it's double the income. And that's the sort of thing that marketers and businesses want to see. So, you know, personalization is everything and the way that we personalize is, yes, we allow people to convey what it is that they want, but also with AI, you can start learning through purchase history.
Luke:
Amazing. So what you're talking about is AI is personalized and has memory. It has memory so that it can give you predictive, it can predict what you want and make predictive suggestions to you. And ultimately with all of that, you're optimizing for time, aren't you? You're optimizing for my time as a user who is looking for a specific thing, for a need that I have.
Yasmin:
Exactly, exactly. You know, I really do imagine this world where I can just pick up my phone and say, I've got an event to go to in a couple of days. It's a networking event. It's casual. It's formal. I don't have anything to wear. And the AI going, how about this? And I know you're going to love this. Or maybe the AI can even look into my existing purchases and say, well, I know you own this, and this, so how about you pair it with this? And that's a whole new app. I am no longer spending all of the time clicking, scrolling, searching around. I'm optimizing and utilizing what I have. I'm buying those hero pieces to uplift my wardrobe, make it current for a new season. There's all these things that you're unlocking now with the AI's capability.
Luke:
Yasmin, we're coming towards the end and everything you've shared has been super insight into AI, into shopping, into other areas. People are listening here and they're thinking, okay, what valuable lessons and insights does Yasmin and Sociate have that I could perhaps apply into other industries, number one. And number two, what skills are really important for the next five, 10 years of a career in this space?
Yasmin:
When we started Sociate and we said we were going to create a curious AI that could see and speak, people said, don't be silly, you know, that's impossible. You can't, you can't do that. What I want to say is that actually with the advances in AI, with what's happening in academia, that barrier of you can't physically because the tools aren't there, it's gone or it's going. And the only way you can advance technology is by really pushing and looking into the realms of possibility and science fiction. So if you have ideas, don't listen to people that say that's not possible. Find out for yourself. I would really recommend, especially if you are on your way to university, to get involved in societies where you can get exposure to the people that are also in research, in deep tech, in academia, and talk to them. And sometimes it might just sound incredibly academic, but try to get to... the crux of what it is that they are working on. And I would say work backwards. So, you know, the general wisdom is find a problem and then the solution. But now with where technology is, imagine a solution and think about where the world might be in four or five years. Imagine and build for that because everything we know and how it works is gonna change. And none of us really have a grasp of what the world is going to look like in five or ten years, really, but we do know it's going to change. So just let yourself imagine and build for that world.
Luke:
I love that. So what you're saying is actually, forget the skills, be curious, have a big imagination and go for it. Be bold.
Yasmin:
Curiosity is the most valuable tool that we have as humans. It's why we decided to give our AI curiosity. Because if you've got curiosity, you can learn, you can uncover, you go down these paths. And none of those paths, even if they do come to a dead end, they're not wasted journeys because you've learnt so much and then you just kind of, yeah, branch off a little bit. So, yeah, you know, imagine, imagine what the world is going to be and let that run wild and it will all be in the realms of possibility.
Luke:
I love it. I love it. And I completely agree with you around curiosity. Be curious about the problem, why it exists, unpack it, and then be curious about potential solutions to it and why they haven't been done yet. Maybe there's an opportunity. Yasmin, in the moment that we've got left and thank you so much for this really fascinating conversation, let's end by sharing with us the one thing that you're really excited about with the future of AI in the next, you name the time frame.
Yasmin:
So to me, the exciting thing is that AI really does a lot of potential in humans. And I think that's how you see it. So, whatever your passion is, if you have ever felt that you don't necessarily have the technical skills, that goes away. And to me, it's that democratization of ability. It's, you know, reducing the barriers of access for so long, especially when it comes to entrepreneurship, you need to have the right networks, you need to have the right friends, you need to have, you know, know how to do marketing, all of that. Now, so much of it is actually within your control and you can execute. And I think that is the most exciting thing for every single one of us.
Luke:
Fantastic, I completely agree. And I know people are gonna want to connect with you, Yasmin, and read more of your content and see more of your videos where you're sharing your knowledge and your insights from Sochiate. Where can people continue to engage with you after this session?
Yasmin:
So find me on LinkedIn. It's Yasmin Topia. If you're not on LinkedIn, I am on Instagram @Sociate AI, or you can also find me on Twitter. I'm not very active on Twitter, but on Instagram or on LinkedIn would be the best place.
Luke:
And on your website.
Yasmin:
And on my website, yep, you can kind of click that button for a demo and at the moment that email comes straight to me.
Luke:
Very good. Thank you so much, Yasmin. That's been a really fascinating session. You've shared so much insight, not only about AI, but about shopping and about the skills needed to come into this space. I'm sure people are gonna wanna connect with you. So thanks for sharing your details there. And of course, thank you very much and good luck with the future of Sociate and saving us all time, finding the relevant things that are gonna help us look amazing or solve problems that we have on our daily lives.
Yasmin:
No thank you very much for inviting me. It's been a pleasure.
Luke:
Thank you so much. Take care.
Yasmin:
Bye.
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur.
Block quote
Ordered list
Unordered list
Bold text
Emphasis
Superscript
Subscript
