How to build transparent and effective AI solutions for insurance?

Oleg

Hi, everybody! Welcome to Devico Breakfast Bar. Here we speak with different people involved in the business landscape, share their experience, delve into the latest tech trends, and explore the ins and outs of IT outsourcing. I'm Oleg Sadikov, and today I'm excited to have David Schapiro, a visionary, a veteran director, and CEO with over 35 years of experience with data analytics and InsureTech companies. Don't forget to subscribe and hit the notification bell so you don't miss a new episode. David, first of all, nice to meet you. Thanks for joining. Could you please start by telling us a bit about yourself and your professional background?

David

First of all, Oleg, thank you for having me. It's a real privilege to participate in one of these sessions. I've been following it now for a little bit. And the sessions are interesting, and it's an honor to be speaking aside so many people that were on this podcast. So, my name is David Schapiro. I've been now, it's already 40 years in the wonderful world of advanced data analytics, software for enterprise organizations. My undergraduate and graduate studies are in mathematics and computer science. I began working in the 1980s as a software engineer, as a programmer. I wasn't a good enough programmer to stay in the programming side. And so, I moved over to the dark side: sales, business development, management, and all that kind of stuff. And after 20 or some odd years, I was taken into the role of being a CEO. And today, as Oleg mentioned, I'm a director, I'm an advisor, and co-founder in companies. Of my 40 years in advanced software, the last 20 have all been in insurance, meaning I live, if you want, I work in the intersection of insurance and advanced technology.

So, I started this when I stumbled upon being the CEO of Earnix in 2006. I had left ClickSoftware. I was a very senior executive, executive vice president of a publicly traded company. We had an IPO. Then I thought I knew everything, and I became the CEO of a small company, Earnix, with about 10, 15 people, not doing a lot of money. Maybe a million a year. It was in insurance, and I thought as an experienced software engineer, experienced sales executive, and business development, I very easily would get in and understand the business. Well, it took me about three or four years to understand what I don't understand in insurance. And ever since those first three or four years, which is now almost 15 years, I've been a student of the wonderful world of insurance, and I've fallen in love with it. So, at Earnix, what we provided, and the company is still around, doing much larger than when I left it. I left after 10 years in 2017. The company offers very advanced pricing, sophisticated pricing for property and casualty insurance providers, large ticket items, enterprise systems, very, very advanced software and data analytics, and optimization, selling to the C-level business people of insurance companies. This can be chief actuaries, CEOs of insurance companies, which tend to be really, really smart people, very mathematical, very analytical.

Many of them have PhDs in mathematics, economics, or physics. And they look at the world by assessing the risk of what they're underwriting. And, I have a mathematic background. I couldn't hold a candle next to them in mathematics, but I enjoy mathematics. I come from a family, even though I was born in the United States, and I've lived most of my life here in Israel, but my father's German background, so I'm very, very square, if you want, and very risk-averse at heart, even though I'm an entrepreneur. And so, I actually understand how these insurance people look at the world. And on the other hand, insurance is underserved by technology. The software systems being run in insurance companies are very, very old. And there's a lot of room to provide additional technology, and there's not really that much technology going in because it takes such a long time to have traction in insurance. Usually, the venture money gets bored after a while and leaves. Anyways, after serving 10 years as the CEO of Earnix, in early 2017, company was already close to a hundred people, doing like 20 times as much when I joined, as far as revenue is concerned, I left, and ever since I've been splitting my time between business and pleasure. I'm happily married for 40 years. I have three children and six grandchildren. So, I spend a lot of time with them.

Also, I have a lot of hobbies. I surf and paddle surf. I sail different types of boats. I do some volunteering work. And usually, I get in quite a lot of skiing, a lot of office skiing. So, I was able to do all of that. But I doubled down on InsureTech or the integration of insurance and technology, again. And I co-founded a company called Planck, where I'm sitting right now – if Oleg would have let me open up, you'd see the beautiful skyline of Tel Aviv here behind me – and which was founded by two very, very smart individuals, Elad Tsur and Amir Cohen. I met them after I left Earnix, after they had already founded the company, but they let me come on board as a third employee, as a co-founder. I call myself an honorary co-founder because I brought, if you want, the experience from being in the world of insurance and, you know, already have a long track record. Both Elad and Amir are like 20 years younger than me, even more. So, this is what I primarily do. I'm a co-founder here. We offer very advanced data analytics and insights to commercial insurance companies in the United States, in Europe, in Australia, in Japan. In addition to that, I am a director and advisor to either insurance companies using advanced technology or companies providing advanced technology to insurance company.

I was for many years – five years – a non-executive director and chairman of the audit and risk committee of ManyPets, which is quite a substantial pet insurer in the UK and the U.S. Again, I joined when they were small, now they're a large company. I was a director at a company called Ahoy! Insurance, which does recreational marine insuring yachts. That's close to my heart. And I'm an advisor to a pricing sophistication company called Quantee, which is out of the UK and Poland. And I'm the advisory board of a company called Herald in the United States, which is an API hub sitting between commercial insurers and brokers. And Sayata, which is a very interesting company with both analytics and a hub for commercial insurance. Besides that, I try to keep up to speed with my business friends who manage insurance companies. They used to be my customers, today they're my friends. If I want, I try to be as much in the loop what's happening in large insurance companies. I try to be in touch with many new insurance startups that are offering new things, new technology in the world of insurance. And also to the venture side, my friends are venture capitalists who fund the kind of places where I work. So, Oleg, I apologize for the long answer.

Oleg

I have only one question. Do we live in the same physical world? You still have 24 hours in a day, right?

David

First of all, first of all, yes, we do. I also, by the way, I like quantum mechanics. We can talk about that. I don't understand it, but I like it. But we do live in the same 24 hours, and I don't live in parallel worlds. First of all, I enjoy what I'm doing. I only do what I like doing. And also, the order of magnitude of stress and effort on a CEO, and I've been there, is by order of magnitude less than anything else I'm doing now. Even though, you know, it sounds like a lot – I'm working, you know, five or six different companies and all these hobbies – but the stress or the workload is much less than I had before. So, it's truly, I only do what I like doing. And so, I enjoy it. Even though, you can see I've lost a lot of hair. It's not always fun, but...

Oleg

Thanks for the detailed answer. What inspired you to found Planck? And could you share a little about the journey from conceptualizing the idea to its realization, implementation, as the leading provider of AI-based data analytics for the insurance company?

David

That's an excellent question. I think I was very lucky in the context of Planck. I joined Earnix when the company already had existed, even though I knew the founders. And when the company was founded, the concept was founded by insurance people, by a combination of an insurance executive, a CEO of an insurance company, and a very, very talented statistician. And the idea was to improve the sophistication of insurance pricing. And since the pricing or the loss ratios are actually the gross margins, when you manufacture an insurance product, you're basically doing the mathematics of assessing what is the risk of the person issuing a claim. Okay? And so, I knew about what they were founding. It wasn't my idea, but it seemed so right because pricing, mathematics, and then insurance, where the product is the price, more or less. And I ran there for 10 years, and it was extremely interesting. And I got very, very involved with our customers. Again, the customers are very close, stayed with us for a long time – you know, many, many years, even now today – decades paying us a lot of money every year because they were getting a lot of value. And what we were doing was we were helping them with very advanced statistics based on the data they have about their customers.

And one of the biggest challenges was the data is not very good. Insurance companies have an excessive amount of data, probably more than any other industry in the world. But the data is not robust. I call it dirty. It's not consistent. It's very difficult to access these many insurance companies when they have grown by acquisition, but they have not integrated their computer systems. So, let's say they sell auto insurance and home insurance, and business insurance are all running on different systems. So, you're insuring somebody's home and car, but they're not in the same database, so you don't know about them. And if they do, then when they renew, since literally the policy admin system could have been written 30 years ago, it's very difficult to extract the data when you're adding. So, the analytics or the pricing is only as good as the data is on the bottom, right? And so we were, if you want, challenged or limited by how good our algorithms would work by the data coming in. So, fast-forward 10 years I leave Earnix, in February 2017, traveled a bit with my wife, and things like that. But in the summer – it must have been in July – I was introduced by a mutual friend, also a technology insurance CEO, to these two young entrepreneurs that founded a company called Planck. By the way, Planck is named after Max Planck, the Planck constant. It's all about quantum mechanics, the resolution of the universe, if you want, but smallest entity that exists. In any event, what they were doing – they came. At that time, six or seven years ago, they were in their early thirties, but they have been working for close to 10 years. One of them, Elad, the CEO, had completed his bachelor of science in computer science before he finished high school and was then in the Israeli Ministry of Defense, where he did very advanced algorithms, AI, working on data about businesses, and organizations, and people.

And when he left the service after about 10 years, he founded a company called Blue Tail, which was all about data about businesses, and they were acquired by Salesforce. And today Blue Tail is the key component of Salesforce Einstein, and a lot of those three years at Salesforce as chief architect of Einstein.

And when he left, he wanted to attack the same problem. His father is a very prominent insurance executive, and he wants to attack the same problem in the world of insurance. So, actually, they found that Planck all about providing data, or it's not data – it's insights. It's what the data is built on, taking more and more data, aggregating it, and combining from different sources, and vision, and natural language processing, and much, much more, but providing data to an insurance company about underwriting a business. Here I was coming from a decade of being very frustrated about the data, which isn't good enough, or it's good, but it's not that if it could be a little bit better. All these outpricing algorithms, underwriting algorithms could be much more efficient, and everything would be good for the consumers, for the customers, and for the insurance companies. And here someone is founding a company like that, so it's just like my dream. And it still is. And I've never looked back. So, the idea was not mine. The idea was Elad's and Amir's, but it just fit where my frustration was about data. There's so much data, but it's not clean enough, not aggregated enough to really enable these great insurance companies to perform their algorithms and statistics on it. So, that's how we founded Planck.

Oleg

Great. Looking back on your career journey, is there any specific moment or decision that you feel has had the greatest impact on shaping who you are today?

David

That's a philosophical question. At my ripe old age – I'm 66 years old – I tried to, you know, I kind of... My father was a little bit of boxing when he was in college. I roll with the punches, meaning I'm a big believer of enjoying the ride. And maybe the best decision I ever made was I want to enjoy the ride. It's all about the journey and longevity being part, take the journey and not giving up. If you're not going to give up, you have to enjoy the journey. So, it's enjoying the journey. I'm not here for the end result. I want the end result to be good. I want to, you know, make money or whatever it is, but first of all, I want to enjoy what I'm doing. So, it took me a while, it took me a long time in my career to focus on that. But I think the best thing for my career and also for my quality of life was to enjoy the journey. I do things that I enjoy doing. I look at what I'm doing as a calling or as something that is more than just a job, and I want to continue to do it. I like there's this in tech, in startups exit. I'm not for an exit. Yes, I wanna make money, but if an exit means I can't continue doing what I'm doing because somebody bought my company or I IPO'd it, which means I don't want. So, I think the main thing I can say, and it's not a decision I made, it's whatever I do enjoy what I'm doing, enjoy the journey.

Oleg

And that's the best way to live. And the earlier you understand that, the easier it will be for you at the end.

David

I think that if one thing I could have done, you know, next time I live again, I will start earlier. But another thing is, I have to say, it takes time. You know, let's say I began really enjoying the journey with my late 40s. I'm not sure if I could have done it 20 years earlier, because I guess it wasn't there yet in my mindset. So, it all flows together.

Oleg

As an industry veteran, how do you stay informed about the latest trends and developments in InsureTech? Are there any specific sources or platforms you're reliant on for staying ahead of the curve?

David

That's an excellent question. I think it's maybe the funniest part of what I'm doing. I love researching. You know, in the olden days, we have to go to a library, open an encyclopedia. Today, I get a name, I read, let's say, a newsletter, I see a name of a company or a name of a person, I immediately google it, find it on the web, research about it. If I don't understand it, I will ask our Gemini or GPT to explain it. By the way, I found that GenAI can be very, very wrong, very, very wrong. So, I don't trust, but it helps me sometimes when I'm trying to clarify things. And so, I'm continuously researching, and I enjoy it. I, by the way, not only do what I mentioned – quantum mechanics, or sailing, or surfing, or skiing – I've been always researching things. For instance, I try to stay on the cutting edge of property and casualty insurance, meaning that's all of the insurance lines, except life insurance and annuities. I don't do life and annuities, and only a little bit of health, but so anything about that, primarily in the Western economy. I don't do it in places like, let's say, China or India, where it's all about growth. China maybe less. But for instance, in Western economy, it's more about there's an existing amount of risks that have to be insured.

How do you insure them? There are new things like cyber and ESG, but basically, so I'm focused on insurance in the Western economy, property and casualty insurance, and new technology. So, whenever I read something about that, it can be coming from an insurance company, it can be coming from a startup, it can be coming from McKinsey, or BCG, or from a university. I hear it, I see it, I dive into it. Where do I find things? Newsletters, on the web. You know, I can be meeting people, and they'll mention a name. I'll write the name down, and afterwards I'll look at it. I will mention the name of a company. By the way, I could probably do that all day. And so, it's really, really, really interesting. I am challenged because I prefer reading. I'm not good at listening to podcasts because it can take a long time, right? So, I'm more reader. And by the way, I practice mindfulness. So for instance, when I drive, I'm just driving. I don't listen to podcasts while I drive. I don't listen to the radio. I might listen to music. When I'm washing the dishes, I'm washing the dishes. I am not listening to something, not even music. If I'm working out, or running, or whatever, I'm not listening to music or podcasts.

So, I haven't yet enough gotten into the listening mode, but reading and research. And today, the web and the availability is great. But I don't believe anything that is sent to me. Again, you know, I go search for things. When I don't trust one source, I'll check sources. And definitely, what is pushed to me – I do not believe in, which is really bad today. Things can be pushing you. You can think it's true, and it's not. So, I don't take anything at par value or face value. And just getting back to GenAI, Gemini, BARD, ChatGPT can be extremely wrong, extremely wrong without any trace of that. They can be adamant in what they're saying, and they're extremely wrong. So, I'm also very, very careful with that. But research I like doing, and so I can stay on the cutting edge.

Oleg

Okay, great. Are there any professionals or leaders in your network who inspire you in your professional journey?

David

Many. Many. I always say, you know, I'm an advisor, I'm a consultant, I'm a director, whatever, I probably learn more from the people that I work with than they learn from me. Well, maybe they get as much from me as I get from them, but, you know, they pay me.

Oleg

I thought you will say you learn from them more than they know.

David

By the way, I think that the capability of being open to learn from the people you work with enables you to know much more than each particular item. So, it's like one plus one can equal three. And, I try to be very open-minded. And so, I can mention many, many people you know, the founding team at Planck, Elad and Amir, the people I worked with at Earnix, Sammy and Ruben, people I'm working with at Herold and Sayata, and the people that I haven't worked with, we just discuss. It's, I think, the opportunity, and that's why I'm also a bigot into meeting people in person. I still travel a lot, even after COVID. I like meeting people because I understand what's happening and try to understand the problem. And so, I say everybody that I meet, I try to learn from. But if I would have to say one person, I would like to mention Alan Bauer.

Bauer is spelled B A U E R, first name – Alan. Alan I had the opportunity of stumbling upon through someone in McKinsey, Conrad, who made the introduction. It must have been in 2007. I was a very young CEO, very early-stage CEO at Earnix. Alan had just gone into retirement. Alan was the president of Progressive Direct in the United States from about 1980 until 2006. You can look him up, he's what turned Progressive into what it is today. Alan is one of the smartest people I know. He's even older than me, probably five years older than me. You know, very prominent background. He has an MBA from Chicago. But he knows insurance. He lived in insurance. He's very analytical. But he was the first person in the 1990s that brought Progressive to sell insurance over the web. And I had the honor of spending a lot of time with Alan. He visited us in Israel. I slept over at his house in Marin County, north of San Francisco. And we're in touch to this day. So, if there's one person, I would say it's Alan Bauer.

Oleg

In your opinion, what are the key trends shaping the future of the insurance industry, and how do you see AI and data analytics continue to impact it?

David

First, I think there's a huge, huge opportunity for data analytics and GenAI in insurance once again because the product is mathematics, okay? So, the better you get the mathematics, the better you're manufacturing your product. It's all about assessing the risk, and if you can assess the risk better with machine learning, GenAI, whatever it is, then the standard statistics, you're going to be assessing the risk better, you're going to be able to price the risk better, more people can be insured, and everybody is happy. That is one thing. And that's why I see the huge opportunity there. The major challenge, there are two major challenges, one is technology in insurance is very old. The software itself, the systems themselves, have not been upgraded for many, many different reasons. There's a different mindset. In insurance, things take decades to build credibility. And when you're funding tech startups, you want things to take five years max. It just doesn't coincide. So, it's difficult to get the venture funding with the patience for the results in insurance. And so, that's challenge number one.

Challenge number, and that's for any kind of technology, challenge number two is pricing in insurance. Insurance is a regulated market. Everything has to be transparent. You have to know exactly how and why you're pricing, what are the parameters you're taking, because otherwise you're regulated. You cannot price based on religion or something, many places based on gender, or based on location, even sometimes. Some places you can use credit scores, other places you can't. So, you have to know how you're building your price. In basic statistics, and GLM models, and things like that, you know exactly the parameters you're using. But when machine learning comes in, or generative AI, almost by definition, it is generating things, or learning to do things, and you don't really know based on what. So, potentially, it could have found a roundabout way to price insurance based on gender, even though you're not allowed to price on gender in the European Union and the United Kingdom, or to price based on a credit score in California, and you're not allowed to price based on a credit score in California.

And so, this transparency is the challenge. Now the challenge is to the technology providers like myself. We have to provide a transparent way, even if it means the algorithms are not going to be good. That's why at Planck we provide a lot of reasoning, explaining how the result was found literally and tracking and enabling to have a chat so you can understand where it came from. Also, the regulators will need to work with the technology providers because you have to give a little bit more room and guidance and just saying, 'Show me exactly the parameters.’ I think that is the primary challenge. Adding to that, in the United States, there is an insurance regulator for every state, meaning there might be 350 million people in the United States, but there are 50 regulators for every state, which means if you're an insurance company covering the United States, you have 50 regulators to work with. United Kingdom, there might be 70 or 80 million people, but there's one regulator. So, since the U.S. is such a huge market, it's challenged even more. By order of magnitude, more challenging there to work with the regulator.

Oleg

Well, I didn't know this. Good to know. What are some common misconceptions or challenges you have encountered regarding the adoption of AI in insurance, and how do you address them in your role as a leader?

David

It's interesting, I guess came from a meeting earlier today with another company. I worked with some of the most brightest people in the world – physicists really doing fantastic things in insurance. And what happens sometimes is non-insurance people think these insurance people are just move so slow, and they don't get it. I can improve your profitability by, you know, 10 million dollars a year, 10 million euros, or 10 million pounds a year by using this algorithm, and you, Mr. Insurance, you want to test it now for three years? You're leaving money on the table. Just do it. Well, that misconception, okay? By us, technology people, there's a very good reason why the insurer cannot adopt technology, because you need actuarial history for a number of years. You don't know your loss ratios really until a year after the policy has been signed or until there could be a tail on the policy, can be two years. There's reasons behind that.

The insurance company cannot default. They have to pass, they have to have enough capital to pass. We, the tech, there's a challenge for us to be able to understand that the decision process is very different in insurance. It takes longer. They need more proof. On the other hand, once they're a customer, once you're providing them value, they're yours forever. Literally, I have customers that have been working with companies for 20 years, paying a lot of money every year, and making a lot of money off of what they're doing, and for the better of all. But it might have taken two or three years to get in. So, the first challenge is, for us, technology providers, to understand that mindset of the insurance, which is the biggest challenge because many of us don't have the patience. Why should I work 10 years in this crazy insurance that I'll do marketing or advertising technology or human resources, technology, whatever it's something, you know, just do something different. So that's number one. The second thing that challenges on the insurance companies – to understand that sooner or later they cannot continue not adopting the new technology. For instance, when I started working in insurance, close to 20 years ago, the average life of a policy admin core IT system and insurance was about 20 years.

Today, it's almost 40 years because they haven't been replaced. Some have, but don't get me wrong. There're great companies like Guidewire and Duck Creek to do that. Sapiens. But now, it can't be 20 years from now. It's going to be 60 years old. It just won't happen. It can't be that, you know, everything will be using in different ways machine learning, and insurance won't. So, sooner or later, somebody is going to come in and move their cheese. Now it didn't happen with Lemonade, and Next, and Oscar, and all the other great companies that haven't substantiating themselves as major players in insurance. Maybe the incumbent insurance companies were scared for a few years, maybe a year or two back, but they're not scared anymore. But they can't be complacent. So, I think the challenge for the insurance, incumbent insurance, my friends who are running insurance businesses, is to understand the need for them also to work with us, technology people, to adopt this new technology.

Oleg

As technology continues to evolve, how do you ensure that your company remain agile and adaptable to future disruptions or innovations? And what steps do you take future-proof your digital information efforts?

David

Extremely difficult. I live in the world of small companies, from one person to, let's say, 100 or 150 people, from no revenue to 30, 40 million. When it gets bigger than that, I get lost. And I think the fact is, you know, why do even companies like Google, innovative company, and Facebook, why do they acquire startups? Because it's the innovator's dilemma. They can't, you cannot maintain. You know, at Google, people leave Google to start new companies. You can't maintain it. So, I think part of the understanding is there's a lot of way to go, and you can lose a lot of hair taking a company from zero to 20, 30, 40 million, right? But, once you're there or at some point along the line, you begin understanding that, okay, either I have to turn this now into a real business that is looking at different things in innovation and be ready to acquire. You know, many companies when they get to a certain side, they have to acquire the new technology. So, the bottom line is you can't, I mean, you can, but the way to do it is different than when you went by acquisition and keeping an open mind. You know, Google is very, very innovative, right? But they still do acquire. And one of the problems is you have customers. And your customer, in my book, customers come first. Nothing is more important to customer. When you don't have customers, you can innovate a lot, right? And when you have customers, you learn from them. But you have to first support them. So, at some point, you can't.

Oleg

Okay. Interesting. How do you navigate talent scarcity in the highly competitive fields of AI and data analytics? And what strategies have you employed to attract and retain top talent?

David

That's a challenge. It's a big challenge, especially today. I'm looking primarily at software developers, data scientists, that area. If someone stays with you for two or three years, it's considered a long time. And where I come from, you know, I work at companies for a decade plus. And in the olden days, 20 years ago, even software developers and engineers had worked five-plus years with a company, even longer. So, it's a major challenge. Let's say, it was worse a year or two ago, when everybody was headhunting talent from other places. Today, it's a little bit quieted down because, you know, the venture funding and things like that. Companies are more challenged, people are staying in place. But still, it's a very, very big challenge. And there are a few things that's not something that can be, I don't see any panacea or golden solution. But number one is you have to be on the cutting edge of the technology, what the developers are using, always, always, always using cutting-edge and the newest tools and using the newest things to offer. Again, it's like your former question, this becomes problematic if you have, let's say, a hundred customers, which are sitting on a few million lines of code, and they've been with you for 10 years. You got to support them. So, that gets difficult, and that's why people leave, you know, maybe the younger companies. But so, one is using the tools as very advanced tools and being in a very advanced market. Again, I like working with startups and innovative companies because you're in a very advanced market, you can attract talent. The company has to be successful, which is not always that easy. And it also has a lot to do with how you compensate your employees. I am a big believer in high degree of equity compensation, meaning if the company is successful, the employee is successful. Unfortunately, today, that's not always enough because, at places like Google, they'll pay them a high salary and a high equity, but that's part of striving to do the game. But it is a material issue, definitely.

Oleg

Okay. Recognizing the challenge of talent scarcity, many companies turn to IT outsourcing as a solution. Have you explored or considered IT outsourcing as a strategy to address specific skill gaps within your companies, and what factors influence your decision in this regard?

David

Very good question. Interesting. I just had lunch with the VP of engineering of Garrett Plank, who was someone who worked at Google for many years. And I think the right way to do it is to have a part of the software outsourced, developed offshore, in different locations. And at a steady state, and by the way, you see this at large companies, you'll see part of byproduct, part of the infrastructure, wherever it is developed, it's outsourced. But the question is how to do it in the smaller company and when. So, I personally, at Earnix, we tried to do it, and finally we had a VP of engineering who had come from a large company that had outsourced, and he was able to get something going. Today, when we try to do things like that, challenge is in order to outsource something, it has to be very clear what's being done, and you have to know how to do it yourself, or something that is like off the shelf. And so, you first have to get to that stage or take something off the shelf. You know, we haven't gotten there, for instance, yet at Planck, but one of the companies I work with, which is even smaller than Planck, from day one, part of their development was outsourced. And so, the strategy, the architecture of the product, if you want – this component – we can define it, and it can be written there, wherever that other place is. This component we're going to be writing. And from day one, they did like that, and they've never looked back, and it's extremely successful. So yes, I'm a big proponent of this. Have I personally been able to successfully use it? Not yet.

Oleg

Got it. Yeah, there are different approaches in terms of IT outsourcing. You can start from the beginning of your journey, and you can add. It's not only about outsourcing the solution, you can bring external experts who will work closely with your team, and all together – your in-house team, their team – they can work as one mechanism to build the solution that would resonate in the future.

David

That is very, very true. And I would not relate that point, and you're correct. We have used external, it's not, it's external expertise, okay? I don't call it, it's not really outsourcing, it's sort of insourcing. They came, for instance, we did not have very good automated testing in one of my companies. We tried to do it ourselves. It didn't work. So, we brought in an external organization that basically helped us build it and then moved on. Actually, we hired one of the people there to stay with us. And so, that is a very, very good point. And this is a good example because we were all about developing the software. You know, engineers can do the testing, and after a while – no. You have to automate it, and especially when, you know, regression tests, that kind of stuff. And so, that whole play we did with external resources, you know, getting our team up to speed.

Oleg

Yeah, yeah, that's the entire idea. In your opinion, what are the main advantages and disadvantages of IT outsourcing, and how do you navigate these factors to maximize the benefits while mitigating potential risks?

David

I'm a big believer in integrating the core and outsourcing the context. For instance, if my core technology is very advanced machine learning algorithms, and that's the core competency of my company, that will do in-house. But you can't sell a machine learning algorithm on its own. There has to be an envelope around it and has to integrate with other things. And so, all of that stuff could be outsourced. Okay, it's not that easy. Conceptually, it could be outsourced, not always because the stuff that is closer to the core, probably couldn't, but if our expertise is we are data scientists, okay, or we know how to run, develop GenAI models, or whatever it is, let's just do that. The problem is you can't just sell that. So, you have to integrate around it. It's sort of like at the semiconductor world. Today, a chip is designed. Okay, a semiconductor chip is designed by very advanced, intelligent engineers someplace, and that's it. They outsource everything. It is entirely developed and manufactured someplace else, which requires a lot of expertise. So, it's different in software, but to some degree, that's the same. What is my core competency? That is what we should be doing, that we should invest in. And we should continually strive to find things that can be taken off of our back and done by experts that do that. If it's APIs integration, I don't know what, testing.

Oleg

How do you foster collaboration and synergy between in-house teams and external IT outsourcing, again, based on your experience? You can share your experience with the quality assurance partner. And what strategies to employ to ensure effective communication, project management, knowledge transfer across different stakeholders?

David

I think this is like when you asked me my role models. Here is where I think one of the things – you know, I'm 66 years old, but I'm still just a young kid with a dream – I want to be able to do just that. I want to be able to collaborate better between the teams in my organization, you know, between integration and R&D and sales and R&D. It's a mess. By the way, at our company, you know, our U.S. team, and Israeli team, and UK team. It's, they all work for the same company. It's not easy. So, that kind of integration is what I try to do better. I have a long way to go. And it's the same thing if we have a developing team in Tel Aviv and in Kyiv, or in Tel Aviv and in Warsaw. It doesn't matter. Or in Mumbai. It's the same thing. And so, I'm still challenged to integrate between sales and engineering. And I'm both an engineer and a sales guy. Same thing here. I can say, you know, what was successful. In the cases where we were successful, was very clean, clear definitions of what is delivered from here, and what is delivered from here. Also, a lot of joint work, you know, meeting each other. Today, nobody works from the office, right? So, what difference does it make if you're commuting, well not once a week, but once a month to another city, which is a few hours flight, might be in a different country? So, I think that finding that way to do it. Also, I've worked at companies, by the way, I was a director, I wasn't an executive manager as a director. But the whole, the engineering team, part of them were in the same country, but we're literally five hours away, and it worked extremely well. Actually, they were acquired. First, they were like an outsource, and then they acquired the company. So, it definitely happened, definitely. It definitely works. I'm just not, I have not achieved that skill yet, but I'll follow you all again sooner or later, I'll learn how to do it.

Oleg

Thanks, David. Thanks for your time. Amazing episode. Thanks for joining me. You're a very interesting, experienced, and active person. I do believe that there are more than 24 hours in your day, for sure. Yeah, thanks for joining. If you enjoyed our discussion and want to stay updated on future episodes, don't forget to subscribe and hit the notification bell. That way, you will not miss on our latest insights and conversations from Devico Breakfast Bar. See you in a week.