Navigating Ethical AI in Business Intelligence

Ever wondered how a seasoned Microsoft veteran transitions into leading a cutting-edge analytics company? Get ready to uncover the fascinating journey of Ariel, CEO of Sisense, as he shares insights from his 21-year tenure at Microsoft, where he played a pivotal role in developing Power BI and Dynamics 365 Sales. Tune in to learn how Sisense is revolutionizing the way developers, analysts, and product managers embed analytics into applications, making data more accessible and actionable. Ariel’s rich background in business analytics provides a compelling narrative on leveraging AI and Generative AI to drive decision-making and productivity, with real-world examples from healthcare and retail that depict the transformative potential of AI-driven analytics.

But it’s not just about innovation; ethical considerations take center stage as well. We tackle the integration challenges of Business Intelligence (BI) into applications, discussing security, context-building, and usability for non-data experts. Ariel emphasizes the importance of seamless BI integration in the development lifecycle and the crucial role of the semantic model in ensuring trustworthy AI insights. The conversation dives deep into the ethical and privacy implications of AI, echoing themes from Orwellian literature and highlighting the need for stringent data privacy laws and ethical guidelines. From handling algorithmic biases to showcasing the versatile impact of AI in industries like sports analytics, this episode is packed with thought-provoking discussions that underscore the necessity of responsible AI development.

More at https://linktr.ee/EvanKirstel

Transcript

Speaker 1: 0:05

Hey everyone, exciting topic today around the power of AI in analytics, an interesting topic for us tech geeks. Ariel with Sisense, how are you?

Speaker 2: 0:16

Good, good, I'm doing great and thanks for having me today.

Speaker 1: 0:20

Well, thanks for being here, Really intrigued by your mission at Sisense and your personal biography. Maybe let's start with that. Tell us about your journey from Microsoft, 20 plus years, to becoming the CEO at Sisense.

Speaker 2: 0:41

I went to Systance. I've been with Microsoft for 21 years, starting from being a developer Some were back late in the 90s and then growing the company to both engineering product roles. Later on I took some executive roles like running Windows Defender, and then in the last I would say, seven, eight years, I took more of business analytics and you know business intelligence-related protocols, being one of the fathers of Power BI, one of the you know great successors. You know SaaS stories, enterprise SaaS stories of Microsoft, so very proud to be, you know, the first four years of that journey and later also running Dynamics 365 Sales. You know making that an intelligent service for sellers and CRM.

Speaker 2: 1:36

So definitely a very long and fulfilling career At some point in my life, I thought that it would be interesting and rewarding also to get some more of a startup experience. That brought me eventually to ScytheSense, which is in the same analytics space. So, from a domain perspective, that was a domain that I was very well versed on, and I saw a lot of potential with what Sisense is doing in the sense of, you know, embedded analytics, and I thought that there is so much future, you know ahead of it. So I joined about two years ago as the CPO of the company and about a year ago I was offered to take the CEO role, and that's what I'm doing since.

Speaker 1: 2:23

Fantastic journey and fast forward to today. What does Sisense do exactly? How do you explain it to folks who may not have heard of you?

Speaker 2: 2:32

Yeah, sisense is basically, you know, part of you know business analytics realm, which is pretty broad and includes both BI companies as well as other companies, includes both BI companies as well as other companies. What we do specifically is really focusing on creators of analytics within applications. It could be developers, could be analysts, could be product managers. Everyone is part of building applications to end users and what we're doing is creating innovation to bring that to the hands. I mean empowering basically developers or creators to bring analytics experiences or analytics applications into the hands of end users in a very simple way to really allow everyone basically to consume data effectively and also being able to connect between data insight and action, which is really important in modern applications. So think about us as the Intel Insight, so to speak, within applications for analytics.

Speaker 1: 3:40

That's fantastic. Well, I need that. My clients are always asking for more data, more analytics, more insights into the videos and podcasts and et cetera that I create. So we'll talk offline, but let's talk about the big impact that AI and Gen AI is having in business intelligence and analytics. It's changing everything, but how? Exactly? What's the impact this year?

Speaker 2: 4:05

everything, but how exactly? What's the impact this year? Yeah, I mean definitely. I mean I would talk maybe more broadly about machine learning, which is still, you know, something that you see a lot you know, when we talk about AI, like taking a problem model that and basically create some AI experience on top of that. I mean churn prediction and prevention comes from, you know, as an example, good example of ML there are so many you know, in each vertical. And then, of course, gen AI.

Speaker 2: 4:34

I mean, when I'm asked, I'm always trying to give an analogy that you know, to me ML is really. You know, if you look at the brain like a human brain, ml is really the logical, you know, left hemisphere of the brain, like a human brain, it may just be the logical, you know, left hemisphere, right, which generates new ideas, you know, imagines possibilities and really creates innovative solutions. And at the end of the day, you really need both hemispheres, you know, brain parts, to work together right To really create a powerful, cohesive and capable, you know, mind. That you know can really tackle both the understanding of the word, which we usually get from Gen AI, but also you want to transform it through insights and actions. So that's really important to understand that there is no Gen AI, which is kind of the new thing that replaces everything we know about AI, all the more let's call it traditional ML models. It's all coming together and I'm super excited about that because we see that as a company, every day is something that really changes how businesses are working and how they're making decisions and how they're driving better productivity. So this is really really important, I think, not just for consumers we all love you know ChatGPT and you know MeJourney and all those tools. It's really going to be important. I think it's still early days, but it's going to be super important also for enterprises, and I can see that through our customers.

Speaker 2: 6:19

You know really playing in every vertical that you can think about, whether that's e-commerce, whether that's, you know, healthcare, where you see ML models analyzing patient data, for example, to create, you know, disease outbreaks, with, interestingly enough, gen AI in the same vertical, creating personalized treatment plans. So here you get it like how the two sides are really creating two very, very different experiences of you know in the same domain. Get it like how the two sides are really creating two very, very different experiences of you know in the same domain. And I can. You know, retail is another good example that we see again and again and again. You know you see a lot of ML, traditional ML, optimizing supply chain management, but then you see Gen AI now crafting personalized shopping experiences right.

Speaker 2: 7:02

Look at Amazon, right. Personalized shopping experiences right. Look at Amazon, right, and how they're, you know, really transforming their shopping experience. So you know I don't know if you noticed recently they added this you know cool system called Ruffles, right, which allows you to basically not just look for you know a product, but also ask follow-up questions. So now this is not about text matching anymore. It's really about like an immersive experience of looking to what you need for you know and then asking questions and really getting you know very quickly to exactly what you need, and I find that's magical. That's exactly how you know all parts of AI are coming together in some very magical way. So I'm excited about you know where we are right now.

Speaker 1: 7:45

Yeah, and it is exciting. However, many companies who aren't Amazon or Apple or Microsoft have challenges in adopting new BI and AI technologies. They don't have the teams of data scientists and developers on the bench, but what are the main challenges they are facing and how do you see yourself making it easier and accessible?

Speaker 2: 8:08

Yeah, that's a great question. I will start, you know, before we get into detail, to basically go six years. You know backwards, right? That's when you know IBM, you know, started with their you know first database and then started also with some interesting idea of how do you create some you know interesting analytics on top of databases. And, you know back, it looked pretty magical. I mean, this is not just about some meaningless data, it's really also about how you create some insights or simple insights on top of that. So BI as we know, that is really here for, I think, at least 60 years, maybe even longer depending on your point of view, but it's definitely a lot of time.

Speaker 2: 8:50

Right Now, what BI technology has really shifted and evolved over the years, right, is the simple, more kind of database kind of AI, going into more of an IT-governed AI and a BI story and then into more self-service BI with Tableau, with other companies, and then, of you know, more self-service BI with Tableau with you know, other, you know companies, and then you know, of course, the cloud.

Speaker 2: 9:13

So Power BI and you know, looker are good examples. But one thing really didn't change for six years and that's, you know, the paradigm of BI, right, and this is really, at the end of the day, very simple. It's about how you get you know like visual face to data through what is called usually dashboards and reports right, so you get some you know visual view could look like a pie chart, you know, combined maybe with some histogram or whatever. You get you know to start to kind of, you know, interact with your data, maybe ask some questions. But it's really about the drill in, drill out right Kind of an experience You're trying to get to you know from a more holistic view of something you know all the way to what you think would be you know, the root cause of you know a problem or an opportunity or whatever. And then you know, you start, you know, to try to find a solution.

Speaker 2: 10:08

That's a very you know it could have. I think in the 90s it made all the sense in the world, but I think the paradigm is really you know, especially you asked me about Gen AI right, about how people are talking with data, and I can tell you, you know, part of our customers are really really not let's call it data literate, right? One company, just to give an example, my end user not the customer, but his customer, which are the end users are truck drivers, right?

Speaker 2: 10:38

So they're using you know the vendor software to basically get you know route for you know delivery of you know some goods or repair or whatever, and they don't really go and care about dashboards or reports. As you can imagine, right, that's pretty complex. So I would say this is like one of the biggest, I think, challenges as you think about how do you make BI more kind of standard, more present in organization. You know, gartner a years ago talked about how organization, even the more modern organization, if you look at information workers, only about 20% to 25% really benefit from BI. And you have to ask yourself why. And when they started to get into the root causes, one of them was exactly that it's really hard to consume that you want to get to the root cause without going through you know filters or you know slicers or you know kind of poking with some. You know strange visuals on your screen. You want to go very deep from.

Speaker 2: 11:44

Here is a problem I'm having, or here's a hypothesis, and I want to go all the way down to you know the root causes or the key drivers. So that's one, I think. The other one is really about where it should be. I live and again that goes back to that kind of you know, I think paradigm that is starting to get in, you know, basically blocking our you know progress forward because, at the end of the day, you know, people don't like dashboards. They really want to be able to interact with data in the context of their work. So, you know, and I learned that, you know, I told you in the beginning I worked on CRM right, crm. You have this, you probably remember this like 90s kind of look and feel of what I call- yes, I still have PTSD and flashbacks from that.

Speaker 2: 12:37

yes, I have PTSD. Exactly, it's a lot about forms of our data. It's a lot about mindset, that my manager wants to know how I'm progressing with my leads and how I'm qualifying and so on and so forth, Instead of doing, you know, work in the context of where I live. And where I live is, you know, on collaboration, like Teams, or you know Zoom or you know whatever, and I have a lot of email which is both a collaboration but also source of data, because I can actually get a lot of valuable signals by just looking at you know that data.

Speaker 2: 13:09

Because I can actually get a lot of valuable signals by just looking at that data. Think about LinkedIn, right now owned by Microsoft. How much amazing networking-related data you can just get from that. So what if I could bring all the data together into one cohesive, non-ed source? And then what if I could also bring experiences into you know what I'm doing right now, as opposed to I'm abandoning everything I'm doing and then context switching to some you know quirky, you know UI that I need to do something with. So I think that makes all the change in the world and what we learned here at SAS, and that's why I'm personally super excited to be here.

Speaker 2: 13:51

The more we can really get into the context of the application, the more analytics will be just part of the application, versus an add-on or kind of a bolt-on experience from the top and just feel part of that application that pops whenever you need to ask something. Then everything is changing and all of a sudden, I love analytics as an end user because it just helped me to do my job better, and I love analytics as an end user because it just helped me to do. You know my job better and I love analytics as a vendor because it's not like a premium, isolated, detached you know kind of module on top of my core. You know action, you know operational module. It's now just part of you know what I'm doing and actually it allows me to get you know more innovation or faster innovation and more, you know, revenues at the end of the day. So the whole perspective about analytics being from a more optional premium kind of a module to just be part of what I'm doing and really connecting data insights and operation or action, that makes all the you know changing the world and what we believe fundamentally here at ScienceSense that modern applications are going to be composite right.

Speaker 2: 14:52

Basically they have modules of either data insight or action and then ways to bring them, you know, tightly together so it doesn't feel like you're kind of bringing something from the top. So that's number two and number three and the last one is the creators. At the end of the day, your you know output or outcomes of analytics is as good as how easy it is to create them right Now, creating a dashboard is easy, but we already said, you know, or I said previously, that this suffers from the old paradigm.

Speaker 2: 15:22

That needs to be changed right. So maybe dashboard is not the kind of default experience going forward. So then, you know, developer ask themselves so how do I, you know, create that? So the one way that everyone is using, you know, today, bi to get them to applications is called embedded.

Speaker 1: 15:38

BI.

Speaker 2: 15:39

It's usually a very simplistic iframe, you know, driven way to just, you know, bring BI from the top. It doesn't really look like the rest of your application. If you saw that, you know, in the past, it looks different, it looks like a bolt-on. It suffers from security problems. It suffers from, you know, not being aware of the host application. So there is no really conversation happening or context building between you know, the host and your BI and then it just becomes, you know, foreign what we're doing in SIS. And then it just becomes foreign what we're doing in SIS. And that's what our mission in life is to make that just part of the development lifecycle.

Speaker 2: 16:18

So we're trying to take away everything that we can in terms of hurdles from the developer that is not necessarily a data guru to really build in a very simple and democratic way. You know those experiences, so every developer can build that into application. So those are the three things that we're really trying to, you know, fix fundamentally and innovate with.

Speaker 1: 16:42

Fantastic and I'm just looking at your website here. Really, you know interesting customers and partners and some amazing screenshots here. Tell us about some of the minefields that you're helping navigate, because there are ethical issues behind. You know some of the AI tools. There's privacy and data protection requirements.

Speaker 2: 17:05

I mean, how do you think of all those?

Speaker 1: 17:06

things behind the scenes.

Speaker 2: 17:08

Yeah, if you look at those minefields, I would basically say about, I mean, talk about two categories. One category is really about is that even helping? I mean, forget privacy, I don't care about privacy. Is that really helping me? I mean, is that making the job? I keep hearing about hallucinations everywhere. Now, that could be a fun thing. You know, when it kind of you know, gets into a teenager mode. You know, when I'm asking the question as a consumer, it can cost me with my job if I'm actually taking that blindly and just presenting my both. You know what are the key churn drivers. And he looks at me and said you don't even know what you're talking about because the data says otherwise. And and then you know, um, obviously you stop. You know, really, uh, trusting that.

Speaker 2: 17:53

So I would say that's the number one thing. That is really. You know, at the forefront you know forefront of every kind of ai installation or ai deployment that I talk with customers. I mean sure that this is helpful and the way that we're trying to solve that, this is actually a real concern. I mean, this is not something you can fix with slides. You have to fix that with a product. The way we're doing that and the reason why companies like Sisense are actually in a really good spot to really solve that is because we have one of these.

Speaker 2: 18:23

You know, kind of you know magic, you know like the secret sauce right that we're having here, which is really what we call the semantic model. And without really going into that, that's a pretty deep topic and we don't have time right now for that. Think about that as a layer. Basically that is built by a BI. You know software that is really abstracting data to analytical queries, so it has all the business terms that you're using and then it translates that to your data artifacts. So when I ask about who are my top customers, it's really understanding what customer means, right, so it can actually drive the right query into the data. So it can actually drive the right query into the data. So the reason why this is so important because that's also almost a dictionary plus, like a system of truth, basically that says what are the relationship between all your entities and what are the facts, what is like the universal truths about your data. Now, once you have that, you can now start to combine between going to the word to ask a question but really grounding that with your you know those entities and the semantic model. You know understanding of your actual, you know ground truths of your enterprise and when you bring it together in some you know, innovative way that's obviously what we're doing here at SciSense then you get like a grounding truth and also a better explainability why I'm answering the way I'm answering. So this is really really important from a trust perspective.

Speaker 2: 20:08

Okay, so that's one. But you asked me about privacy. Right Now I have to tell you ethical AI and privacy. Funny enough, this is something I'm thinking a lot about. Almost from my geeky childhood being a teen that reads too much Asimov and then watches science fiction movies I always was fascinated about AI, robotics and what's going to be the way that they're going to influence us in a way that makes sense for humanity, not just the way that helps humanity, right. So let me throw some thoughts there, and obviously this is like could be almost a conversation of its own right. It's such a deep and interesting topic. So what is the dilemma about data? You know, privacy and surveillance? I mean just you know we all, or most of I guess your. You know, I mean just you know we all, or most of I guess your, you know podcast audience probably heard or read 1984 by Orwell and definitely worth thinking hard about. If you read that, you're probably terrified initially, and then that led you to think a lot about dilemmas of data privacy and you know surveillance.

Speaker 2: 21:16

So that's one thing that you thing that we need to fix fundamentally, Matrix, which is a pretty cool movie. I liked the first, by the way episode. I liked the sequel. I don't know why they don't stop with the first and the best. They have to continue doing more, but it is such an amazing movie, such a thought-provoking one, and if you think about what it represents, you know from this discussion, it's really about the challenges in our what I would say algorithmic biases and transparency.

Speaker 2: 21:46

Right, you know, you have this. You know, for example, the hidden nature, right, of metrics that represent what's called the black box problem in AI. Right, where you have decision-making processes of AI systems are really opaque. Right, you don't really know how the machines behave and how they think and how you know, and so it's really hard to understand how should I behave to really, you know, satisfy them? So that's really tough. And then you know you look at iRobot, of course, which is one of the most legendary books by Asimov, and that talks about the foundational ethical principles. Right, I mean, those are the three principles. Some good friend actually told me there is a fourth one, but it's really helping you to kind of create frameworks of how to think about the interaction between machines whether that's software or physical robots and humanity.

Speaker 2: 22:43

So I would say, if I look at the future and what needs to be happening, I would say there are, I would say, four, probably areas that I think that humanity and, obviously, government, we need to think hard about. Number one is really about data privacy and governance. I would say you have to implement some, you know, stringent, you know data privacy laws and governance frameworks to, I would say, protect individuals' personal information. Yes, there are PII today. There are some, you know, obviously already governance around that.

Speaker 2: 23:18

I think that's going to have to change to be even, you know, deeper than what we have today. You have to ensure transparency in how data is collected, how it's stored and how it's used to really build trust and prevent misuse. So I think that's number one. Number two, what I said earlier is about algorithmic biases, right? So this is really about developing, you know, a regularly, maybe, audit systems and using diverse and representative data sets to minimize biases as much as possible. We heard about Google and some of their famous biases, you know, in how they kind of label pictures and make some, you know, build some of their models. It's a big deal. Of course, you have to ensure fairness here, and prevents, you know, AI from being perpetuating existing inequalities.

Speaker 2: 24:09

Number three is transparency and explainability. I think we touched already on that when I talked a bit about some of the challenges that we're dealing with, and I think you have to probably look at LLMs but also about enterprise you know kind of universal truths and see how you bring them together. I think that's the only way to really make sure that you get both explainability and also avoid hallucinations in those use cases. And the last one you know, maybe even definitely not the least, maybe the most important one is ethical frameworks and oversight. Right, you really have to adopt ethical frameworks like the Asimov's three laws of robotics I'm not saying that's false 100% from today, but what I like about it it's a framework that is easy now to judge whether a behavior makes sense or not.

Speaker 2: 24:59

And I don't think that today, if you look at, you know, open AI and all those LLMs and the conversation that they provoke, you don't really understand what is the framework like? What is good enough AI right, what is like? You know everyone you know is talking about. You know the ultimate AI, which is going to be like a human. You know great AI. How are we going to really, you know, apply that in a way that is ethical right for humanity, and I think you know everyone talks about it, but I never saw anyone suggesting a framework.

Speaker 2: 25:29

So I hope that one of those LLMs or maybe one of the big you know companies like Microsoft or Google will come with that framework. But I think that the world is really missing that right now. So to me this is for kind of top-of-mind things, but this is so fascinating and I really look forward to see that evolving A really great insight.

Speaker 1: 25:53

So much useful thinking there. Back to Sisense. You have so many amazing partners and customers Just looking on your website. I don't want you to ask to pick a favorite child or anything, but maybe you can share some success stories where someone either in a software industry or outside of software, maybe used analytics to innovate.

Speaker 2: 26:15

So I have 1,500 children, just so you know. So I'm a really proud and busy dad. So let's see if I can maybe talk about categories more than just, you know, individual companies. I can give some examples if you're interested. One of the cool things about this, I think I said initially, we're kind of an Intel inside, so you're powering so many kind of diverse experiences, right, because you're basically powering so many different type of verticals and applications going through digital transformation with AI and BI. So it's really cool to see how many kind of diverse things that you're actually powering. So a few examples. Let's start with sports and analytics. Or sports and analytics, or sports, you know, and athletics Analytics is part of that, right? Usa Swimming I'm sure that you heard about that the national governing body for sports of swimming in the United States. Right, they use CITES to provide interactive analytics to their members.

Speaker 2: 27:16

So the partnership that we're- having with them allows athletes and coaches as well to analyze performance data you know more effectively improving training and competition. So here is what I'm proud about, right, usa Swimming, as you can imagine, evan, doesn't have obviously the same you know budget like NBA or NFL or you know what have you Now NBA and NFL can have armies of those analysts and really get to the you know most kind of you know intimate small detail that helps you to get you know 1% better. But what about the long you know tail of all those you know organizations like USA Swimming?

Speaker 1: 27:58

And how can we?

Speaker 2: 27:59

really support. You know athletes throughout the spectrum, you know and you know with different, you know varying budgets to really be at their best. And that's, I think, where you see analytics, you know and side sense coming. You know beyond the just. You know the one, two, three, you know top players there. So that's actually a very good example that shows you that there is always such a long tail for every industry, you know, when you think about the one, two, three.

Speaker 2: 28:25

you know top leading companies and that's where we really have both. You know business influence, but also you know influence in terms of getting people better. So it really makes me feel good, you know, starting my day thinking that I'm helping such organization. From there let's move, you know, to completely different realm. So NASA, right, aerospace and research right. I mean NASA is our customer one. I mean we're working together and we're collaborating to really leverage AI drivendriven analytics into their operations and you can understand that as their way of building the next spacecrafts or whatever they're building there. Obviously, I don't know exactly what they're doing at every point of time, but it's really so cool to understand that they're using our embedding into their own system. So here is an example of how you basically take a big system that they're using maybe to design something or to build something or maybe even manage some they have a pretty complex supply chain, as you can imagine and then use AI and our analytics to really form better their decisions and, eventually, their actions.

Speaker 2: 29:40

So here's another interesting example. Then you get, you know, to airline operations. So Air Canada is one of our. You know customers and they're using us, you know, basically to optimize flight operations and enhance their passenger experience. So basically, whenever there is like a flight brief before you know on the ground, before they get to the aircraft, you know the flight attendants, you know the pilots everyone is having. You know dashboards and analytics to talk about things like you know weather, you know safety instructions, you know what's going to?

Speaker 2: 30:22

be the journey to destination and what to expect, and this is so cool to know that they're using my software to basically plan their flight and then they're all aligned and you know store that. So here is a completely different example and I'll give you one of the ones that I really like. You know, maybe the most and this is completely out of nowhere, maybe so I'm not, you know, able to share the name of the company, but what they're doing is crisis management and mental health. So this is like an organization, a crisis management organization, and they're using Sisense to support suicide prevention efforts. Sounds scary and specifically, individuals contemplating jumping from the Golden Gate Bridge. So you can imagine bridges you know are not just you know cool places to look at or to drive.

Speaker 2: 31:13

It's unfortunately also a place where people choose to, you know.

Speaker 2: 31:16

At least consider ending their lives unfortunately, and so I don't know if you heard about it, but there is almost, I mean, throughout the bridge. There are places where you have something that looks like a phone, which is, like, you know, kind of a one-in-a-tenth crisis. If you have any suicidal thoughts, you can just pick up the phone and you get some human on the other side. But here is the thing you have to get to a very quick decision-making, you know, given the context of who you're speaking with, and you know what he has in mind. You know, maybe some data that he's, you know, giving to you, if he's, you know, willing to do that. And so there are so many factors.

Speaker 2: 31:57

Even the weather is a factor here, and so what's happening? We accumulate all those data right about, you know, past calls and we are able to really inform in real time the agent on the other side, and we are able to really inform in real time the agent on the other side how to handle best, you know, both in a sensitive way, but also in a way to kind of, you know, drive the right action on the other side to minimize, of course, any attempt to suicide. So I'm really proud about, you know, fundamentally, saving life at the end of the day, and this is such, you know, a great feeling, so here is another place maybe very unexpected where you see Sysense, but this is in a nutshell.

Speaker 2: 32:33

I hope that I give you a taste to how many diverse use cases, both software and outside of software, we're participating in, and there are dozens of verticals like that, which is pretty cool.

Speaker 1: 32:46

Oh, very cool and amazing to hear so many examples beyond traditional SaaS or tech. Yeah, exactly, Really well done. It's also great to see your former employer doing so well One of the or the biggest company in the world maybe right now, depending on the day. Any leadership lessons you took from Microsoft under various leaders, I guess, including Satya and many others what did you take with you?

Speaker 2: 33:15

Yeah, I would say one thing I don't want to come with the list, so I really want to learn, try to learn one thing, which is what Satya, you know, refers to a growth mindset, and it's really about the understanding, especially this crazy, you know fun, scary, you know error that we all live in. That is, at the end of the day, success is all about learning, it's all about adapting and it's all about you know really understanding you know what customer really want and just experimenting and you know getting to. You know what works or doesn't, by really taking a very modest, humble approach and really understanding that at the end of the day, you're not expert in anything. You're, you're always a learner and what I found and I think that's what I'm also inspired, of course, by statia that I mean that's what it brought us, us his kind of style and basic philosophy to how you run the company and on the day we cannot afford especially on a very large corporate that's, you know, saw a lot of success over the years to have this, you know, opposite of growth mindset, which is the fixed mindset, like I know it all.

Speaker 2: 34:26

I know you know how success looks like. You know that's the result when you read, you know books like the Innovator Dilemma that shows how the most amazing companies on earth, you know, stumbled and you know went out of business. This is because they were so trapped in this fixed mindset you know thing. So if there is one thing that you know you learned from Microsoft and I'm taking that also into my personal life, not just you know professionally, it's really understanding that I keep learning and you know, look at the last year I was never a CEO before.

Speaker 2: 34:56

Yes, I was an executive in a large corporate, but trust me, evan, you know and that can serve you know, another discussion of its own. It's so much you know, different universe of being a CEO of, you know, a medium company compared to be an executive, as you know, highest you are and be making an impact as large as you are on a large corporate like Microsoft. There is one thing I really want to you know, your listeners today, to kind of bring home this is really the growth mindset thing and really be open, humble, hungry and really focus on learning, because we learn every day and that's how we succeed.

Speaker 1: 35:36

Well, that's a mic drop moment, so at that point I'll just say thanks for being here Really insightful and informative chat and congrats on all the amazing work you're doing, onwards and upwards.

Speaker 2: 35:46

Yeah, thank you very much and again, thank you so much for hosting me and I really enjoyed this conversation.

Speaker 1: 35:52

Thanks so much, errol. Thanks everyone for watching. Take care, bye-bye.