 Yes, so I have to thank Yana for setting up the discussion so well. She's talked about data-driven innovations and the theme of our panel discussion is basically AI and humanity. And I think that ties in very well with what's happened in the discussion before they talked about examples of how data is being used. But data never actually makes sense until it benefits the end users, the humans behind the technology. So we have here today Charles, who is the director of Accenture AI and Janet, who is lead data scientist at Oracle and also Needy. And Needy is co-founder and CEO of a startup called Podcast. It's very exciting. It's applying smart technology to the shipping and the logistics industry. So let's just dive straight into the discussion. When we talk about AI and humanity, I'm actually going to put a controversial twist to it. So Scott Belsky, who's the chief product officer at Adobe, said within the first 15 seconds, any human is either lazy, vain, or selfish. When you open your Facebook app, you want to see what your friends think about you. When you open the graph app, you are looking to save time with a taxi because you'd rather not walk. You're lazy. So I was wondering, opening it up to our panelists, how do you, when talking with your customers, start to move beyond those human tendencies and build trust with your customer to deliver something that's actually of value? Thanks for the question. I think it's really interesting. I like the topic about human-centered AI because I fundamentally believe that technology is ultimately an enabler and we should not really develop any product, any technology until there's a need from the customer. So as a startup, that's the whole lean startup approach and that's what we followed. For the first six months in building our startup, we did not have any product and we did not really spend time in building the technology. All we did was go out and talk to customers and understand what their problem was and what they really wanted. Only when we had a good idea and we felt comfortable that, hey, this is a problem we are hearing repeatedly and it seems to be real and there's something in it. Then we went out and we started getting LOIs from customers and we said, we can do this. It was just our confidence and the technology that we knew we can build but not having even a prototype we said, let us build this, give us time and funding and we will build this product for you. That's how we started. So I really fundamentally believe that it's important to hear what the customer is saying and only then building something and making technology an enabler rather than sort of, you know, it's not about losing jobs because of it. It's really about changing the kind of jobs we have, the kind of skills we have. All right. So for me, I think I was like smiling because earlier the quote you said is very controversial but I think I will kind of say that humans are not really that mean or vain or, you know, like humans are naturally nicer. I'm more on the positive side. I see, I see. Okay. You're a positive person. Yes. So I think it's being humans, being humans we are kind of have this survival instinct. Right. So what we usually do is that once we are, once we see a threat, then we tend to exhibit those kind of qualities. However, and from the AI perspective, right, going back to the human centered AI, I think once we have shown how technology, how the evolution, how the advancements in technology could actually help humans augment where humans are kind of weak and then it becomes, they become more less defensive. Yeah. Very completely. Yes. So I think, you know, having the principles around AI were first from a technology company. It's actually our role to help customers, help people to understand what really, what AI really is. Because sometimes, you know, maybe media movies all these have kind of, kind of put some negative flavor to AI, like AI will replace jobs, AI will kind of kick you out of the companies and all these, so all these negative stuff, sometimes we just have to, from a technology point of view, we just have to kind of share what technology, how technology can help you. Right. So that's one. And second is actually it's not meant to replace humans, but actually to augment where humans are weak at. It's actually to help you, right, to help you do your jobs, to help you in various aspects of our lives in the community, in the society. So there are many good things that can actually be derived from it. So I guess, you know, like there's so much potential in it, as long as, as you've said, you've centered it around humanity, understand the impact, make sure it's towards the good, then I think there's really a lot to it. Yes. Well, it's hard to add that to that, so I agree with those things, and it's true that one of the first challenges I may have most of the time is that, well, we either have stakeholders understanding, I mean, who have a goal to implement AI, and at the same time, they have a wrong understanding of what is AI. So the first thing when you talk to people, most of the time about AI is they see Terminator. So I think we can have a more realistic kind of threshold, an element of discussion, that's what you were mentioning. So I agree on that. So how you go between the two, how you make sure that, well, AI is the right solution because potentially there is a lot of problems or issue or any kind of challenges who can be tackled and answered by more conventional algorithmic approach still. So that's one of the elements. I think also there is a switch in the industry, a change in the industry. We are not even the process, the way we are developing solutions is not anymore, where we give you a set of requirements and we see you in six months and you deliver, we know with all the agile methodology, you can integrate user into the loop, being at the beginning, being at the middle. Every single sprint you can figure out if you're going to the right direction because at the end, I don't know if to take the code. If we are selfish and so on, I see it in a different way. We want something that makes a difference for us. It's not about selfishness, it's about being efficient. It's about optimization. Yeah, exactly. If I don't need it, why would I open it? So I think that the whole industry is changing an AI project. We were talking about that before, a typical project initially. All kind of technology projects were very linear. We had a set of requirements and you just knew you could even code precisely. You knew exactly how many days it would take you to go there. The problem is what we are doing in AI because you have so many unstructured data, so many emerging of data making sense of many different hidden patterns. It is not that linear anymore. You go through iteration and you try something, sometimes you fail, but you learn something, hopefully. Then you do something else and that's how you get to the objective, to the solution you're going to provide to a consumer hopefully. So I think those are the two. So to sum up our panel's opinions, would it be right to say that building a human-centered AI means always being in touch with your end users, with your customers, and also maybe changing the way you work to becoming a lot more iterative, a lot more oriented around experimentation just to make sure that you're constantly getting that customer feedback? So we've talked about the human side of human-centered AI and I'd like to also touch on the more technical aspect of it. So again, I'm going to bring in another quote from Reid Hoffman, who's the CEO of LinkedIn and he talks about this idea of blitz scaling. I don't know if anyone has heard about that. I'm sure as a startup founder a lot of emphasis now is on hyper growth where you're expanding into as many markets as possible to as many customers as possible. And so Charles and Janet, you both work with very global companies and I know Needy as well as a startup founder, you're very concentrated and focused on growth. How do you manage this idea of hyper scale and growth with issues, very human concerns, like keeping in touch with your customers' needs and making sure that you continue to serve them? Do you have any specific stories or even horror stories, failures that you want to be open about? As a startup, we're still in the initial phase. So just going back to what Reid Hoffman says he talks about, startups going through a life cycle, starting as a family, so a very small team focusing on something specific, moving to a village, a city, a nation. So that's how a startup sort of scales and that's where he talks about sort of growing and scaling much more rapidly. So we're still, I would say, at the initial stage, at the family stage. I think the challenge of growth at an initial stage is really about how do you make sure that you're able to say no to a customer? You're able to keep focused on the product that you're trying to develop because yes, you need to have that ambition that the startup starts the particular customer segment and then scales across customer segments so you build a particular set of product and then you keep adding features and modules to it so you make it grow. So you need to have that ambition but still at the initial level it's really about how do you keep focused and how do you make sure that you're able to make because you have limited budget, you have limited team and resources and time and you want to do the best possible in that and make sure you deliver something great and not just something average across multiple segments and multiple modules of the product. So I think at this level it's really about how do we manage the interest that we're getting from customers but at the same time keep focused and then the second thing related to growth is as the startup grows and as we get more people we start with maybe one or two founders and as we start hiring people and talent how do we make sure that we're still we're thinking not just about the customer growth and the product growth but also the team growth so as founders how do we keep upskilling so that we can manage and we can lead people who join the company and how do we make sure that the people who are joining they integrate and imbibe the culture that we have and they are taking ownership of the product and coming up with new ideas about how the product should be launched and scaled so I think those are sort of the growth challenges that we face at the moment. Maybe an example from the customer perspective also relating to the discussion that we had earlier so about AI it's very hard for a user to just start trusting an AI tool which gives a certain suggestion or a number or a data point when that user has been doing the same work for the last 30, 50 years and saying hey I've done this all my life and now you're saying this technology can tell me something better that I don't know so it's really about while scaling not just giving that AI tool but also enabling the user so what we do is we give the impact of economic patterns on how cargo flows so if there's a US-China trade war how is cargo going to be flowing between China and the US and how is that going to be different than how it was before the trade wars so if we just give a number this is the new demand or cargo flow that will happen between the two trade lanes the user might say how do I trust the system or why do I trust the system but if we give a reason if we add meaning behind the AI and we say this is the amount of cargo we see flowing in the next few weeks between these two countries and these are the reasons why we think this is going to happen because the currency rate is looking like this because there's a typhoon in Hong Kong and so on and if we give those reasons and add meaning to AI that's where the user is more empowered and he starts believing in the AI tool much more so I think while we're growing we need to start thinking about what are those things that we built around the AI product that the users feel more comfortable and want to use the product more so it's about building trust again we come back to that word again funnily enough Janet Yeah, okay I think from a bigger organization perspective the pros and cons in terms of scaling or the challenges and the success stories are kind of different in a bigger organization so I guess in terms of growth in a company like Oracle we are lucky to have access to assets and resources coming from a big company so with all these trends around technological innovation and all these it's easy for a company like us to kind of hire more people even acquire companies like recently, if you have heard about it Oracle has acquired the DataScience.com to kind of augment the skills to augment the strength in terms of data science so it's easy for companies like us to do that easier but at the same time I guess if you go more into the action so how do we actually help customers and so on and so forth there are different set of challenges so for bigger organizations typically we will be organized in different product groups and what does that mean is in terms of solutioning it might be quite tough from a customer point of view because if we talk about artificial intelligence essentially what we are looking at it's not a black boss it's not a magic so what it really is is that AI consists of multiple technological components from data management from analytics from applications basically you can think of all the various components put together to make a machine intelligent so if within a big company like us sometimes these components are being handled by multiple departments so collaboration in terms of scaling solution in terms of building something that is really relevant and useful for the customer could be quite a challenging task so for us who are in a kind of position to influence, to educate our roles are actually very important from an internal perspective as well as from a customer side so it is for us to kind of more than the technology itself it's to push the education push the awareness as to what really the solution around AI is it's not just like a hype and so on it's really putting down the thinking and the impression like 10,000 feet or 100,000 feet down to the ground and say this is how we actually do AI so again like going out and talking and educating people I don't know about Charles as well do you find the same experience so first if we talk in terms of growth and being able to expand your business tremendously from one day to the following day I had created a start-up before and we were quite successful we were doing so we were disbursing loans in the Philippines and we had our own credit scoring because it was targeting the unbanked people I see that in terms of capabilities I mean we have fantastic technology today AWS, Google Cloud obviously Oracle of course so it is very easy to scale on the architecture perspective it is not a problem anymore there is a cost issue so you better be sure that your schema and infrastructure make sense otherwise you may have some big surprises but otherwise it's fine so for example in our cases it's not tremendous numbers but we went from few hundreds applications on one day and the following day we had around 100,000 applications so in all infrastructure not a problem, nothing no load time so that's one thing in terms of technical capabilities not a problem in terms of having models capable to get the personality of everyone to have a model fit on, I don't think it would work I think there is a lot of differentiation that we would need to do so I agree also in terms of explainability even for the stakeholders but I think the issue for me will be or the challenge actually more than the issue will be how you handle differentiation between the human if we could still talk into that human centered perspective is really explainability and how you handle differentiation and what answer you provide to that and that's just a stakeholder answer to provide I think you've touched on a very interesting data science problem actually because you're now talking about many different customer segments how do you optimize the model or essentially the service that you create to deliver what those customers want and actually give that personalized experience going from one model to a lot so I think we have had a very good discussion thank you a lot to the panel I think we've had some interesting themes around making sure that you know what your customer wants working in iterations talking about collaboration and this idea about trust actually becomes ever more important even though we're talking about technology so I would like to thank our panelists for their time and also thank you to the audience for your attention and with that I will hand over thanks a lot