Amir Raskin is Sapiens’s Data and Analytics Product Strategist; he is responsible for the senior customer interactions and roadmap of Sapiens’s Data and Analytics value proposition for its customers worldwide.
Amir brings with him more than 25 years of global experience in the data and analytics domain, from data warehousing and BI to predictive analytics, working with large enterprises in business and technology consulting, product, and delivery. He has also founded two start-ups and was involved in several large-scale insurance data projects.
Two major innovation-related topics are top of mind for commercial insurers worldwide: data analytics and the use of artificial intelligence (AI). Amir Raskin, Data and Analytics Product Strategist at Sapiens, joins Kristina to discuss how companies can uncover insights using capabilities that were once seen to be like science fiction and drive customer service, top-notch experiences, and profits.
KRISTINA PODNAR, HOST: Welcome back to the Power of digital policy. Today I'm excited to welcome Amir Raskin, who is Sapiens Data and Analytics Product Strategist. He's responsible for the senior customer interactions and also the roadmap of Sapiens data and analytics value proposition for customers worldwide. Amir has been in the space for 25 years or so. I'll ask him about that in a moment. But he's worked with large enterprises in business and technology consulting product and in delivery. Amir has also founded two startups and was involved in several large-scale insurance data projects. So Amir, first of all, a big welcome in a thank you for taking the time to hang out today and share your insights on what's honestly the biggest topic of the year and of the decade, I think.
AMIR RASKIN, GUEST: Great. Thank you very much for having me here.
KRISTINA: So did I get it right as 25 years or more than 25 years now that I'm trying to date?
AMIR: It's, in fact, more than 25 years, but it's like purely 25 is in data analytics from the moment this domain has started.
KRISTINA: Amir, you are an expert in digital analytics for Sapiens, headquartered in Israel. We are seeing a lot of great companies and digital innovation coming from the country. Before we jump into the analytics and the insurance industry part of the conversation, can you tell us a little bit about the Israeli startup scene and why it's called the startup nation?
AMIR: Yes, that's correct. And many people have been trying to unlock the secret sauce. I can only add a few of my insights into that. I can demonstrate it a little bit in on myself because many of these Israelis who are excelling in the startup domain as started in the Israeli defense industry, which gets many people knows that it's a fruitful source for a lot of technologies and knowledge. So I'm going to say that first. And secondly, I think there is something with the Israelis, maybe because we are a young nation, and I really see that we are not afraid to try because it's not like we've had stuff for decades. Everything is new and changing. So I think it is the philosophy of life. It's all over, for I have people that I've learned from who will start this, and I am startup business well and people that have graduated. So it's almost natural for people in Israel after they graduate from universities: Where's my startup? Who can I join all what can I initiate?
KRISTINA: It must be an energizing space to be in and right, it's very exciting.
AMIR: I like it. I don't know anything else even though that they've been many as you know, because I've lived in Paris and lived in London and I think there is nothing like this that, that's correct.
KRISTINA: So you have lived in a lot of different places, and you've done some pretty cool things in your career such as forming the information and statistics division at the Central Bank of Israel, the Global Information Center in Teva Pharmaceuticals as well as additional major revolutions is what I'm calling them in some of the largest organizations in Israel and Europe. Do all of these organizations have a common business driver or market opportunity. What is the common thread really between all of them if there is one?
AMIR: I think that the common one is the drive it to change. In any of the major initiatives that I was involved in with large organizations, there was a major change behind it. If it was with Teva or Teva, because the Americans like to call it, it's the largest generic pharmaceutical company in the world, more than fifty percent in the States, Teva has decided to move from an international company into a global company which changed a lot in the perception of the management of what does it mean to manage this giant. They've started I think when I started working with them, they were like 6,000 people, and they became more than 40, and many of the management practices and the expectations of the board has changed dramatically in a relatively short time. So that's just one example.
KRISTINA: And so thinking about that, especially sort of major forces is they're sort of a perfect scenario where organizations should start paying attention to data and predictive analytics? Because you know, one of the things that I often find is that while many folks in the leadership roles, in the CEO role, they're very versed in business. This is great because they have to run a business, but they necessarily don't understand data and predictive analytics. Board of advisors or, you know, the board of directors aren't necessarily versed in data and predictive analytics and really what you can do with it. And so, thinking about whether it's a large multinational or even a small company that's looking to get acquired, as a startup, I'm not sure that everybody really understands the power of data and predictive analytics. How do we get to the point where more people do?
AMIR: That's a good question. If I had the perfect answer, I guess I would be doing much more than I'm doing. I think that, generally speaking, people talk about it, and they don't do enough. I think that what I found, and that's one of the reasons I've chosen Sapiens. It's even doing I'm emphasizing even the insurance industry started to realize that there is something in it and maybe reports that is going to manage that level X and then he's taking the report and doing some changes then sending it to his manager, and then he sending it to his manager. But that's not really the way to change a business. People started to understand that data means value, and I think it will mainly lie in all of the other margins to become very much like, what do you mean by that can be demonstrated by even a functional industry. Let's take the Plastic industry. So if you manufacture plastic, so it is the level of operational efficiency. I guess you can do five percent better. If your company X or company Y. But if you play with data, you can do twice as well. And when you understand that and when you are other alternatives to really do similar amazing stuff with just making your company better with the current best practices, you need to understand that you need to re-engineer a company or to re-engineer the processes in the company. At the moment, you understand that data analytics immediately come into play, together with digital, by the way, and we can talk about it a little bit, but generally speaking, this is the first candidate for, I guess at least 5 to 10 years in the insurance industry, which is one of the latest to the party.
KRISTINA: Do you think they are later than banks, or are they on par with banks?
AMIR: No, not later than banks. I've been working a lot with banks, and definitely the understanding at the senior level, the strategy division, and the board became before. I think the reason for that is that is one of the CEO of an insurance company told me we make the products very complicated so the customers do not need to understand them so that we can do more. It was a bit sarcastic, of course, but there is something into it. I think regulation and a lot of them I would say if fintech that came into the playing on came first to the banking and make them made the product very simple. So, the very little product in the banking industry which are not complicated today. And that's one of the keys. In the insurance industry, the products are still extremely, extremely complicated, complex.
KRISTINA: That's an interesting point actually, because I was thinking as we were getting ready to help on this call today that you and I are both believers in power. I wrote The Power of Digital Policy of book, you speak about the power of data and analytics. And I was thinking about what is really power and to me a lot of power comes out of balancing out risks and opportunities. You just touched upon that. I mean, things like regulations are risks, but that's only if you're not compliant. There's also this other, beautiful part if you will, that's the opportunity, and you know thinking about data and analytics as an opportunity, and as a contrast to some of these risks. So can you kind of helped me understand like how do you really talk to people in the insurance space or in any other space really, about managing, not just the risk, right? Because a lot of these organizations are risk-averse. And so it's all been about risk, risk, risk, but there's also opportunities, right, really crazy good opportunity. So, how do we talk about those and get people to notice?
AMIR: Yes, that's true. Because if you go to the heart of the risk in the insurance area, it's the hardest, I would say, the hardest call to make because they are really into protecting the company. This is the role. Okay, but there are balanced with other creative people in the organization. And usually, we are talking to these guys, and some of them I would say the new wave of analysts, are also joining it. So, when you talk to sales people in the insurance company, when you talk to people who are really into changing dramatically operational efficiency in claims, in policy submission, you see openness to explore new wave or so in new ways. And when I'm saying that, I mean that one of the things we've done at Sapiens has worked with this redesigned the vision of data analytics for the industry as we see, as we implement in our product sweet. We want to make open the processes of the core systems as they are used to call. As people in the insurance industry used to call the operation system, the core systems. We want to make them active. So, we want data analytics to be some kind of insight for an analyst or top-notch C level managers. We wanted to be available for every process and for every process worker. We usually use the term "knowledge worker". This was like ten years ago. I think Gartner started with this term, and we met people that know how to collect a lot of data, and they're very knowledgeable, and they're supported by data. They make an amazing decision. And once they have that this insight coming from somewhere, they will change the company. It's not working. This way, company is being changed by the people that work in the shop flow if I can use this term for the insurance industry and is supported by a range of processes. And when you make the processes active with data analytics, you see people started to sing. I'll give you a use case which is some everybody understands in the insurance industry. People, they get claims? Right? This is like the garbage factory of the insurance; you get a lot of claims, many people are claiming, and then at the end, if you get if you give them the set the same attention to every claim, you're not effective based on data analytics. We tell then, maybe you can resolve 20% of these claims immediately. We can tell you that in most of the cases in this specific strain with all the characteristics. You shouldn't have touched it you you are already going to approve it. We know you're going to put it here. Look, we can test it like a few months ago all over the last year and show you that you always approved it. So, leave it. Just to approve it. What are you doing then? If you really see look into Lemonade, and some other startups (Lemonade is more than the startup today), you will see that this is the way they were building the business from scratch. And insurance people are open to listening to this message because you make it close to the daily reality.
KRISTINA: Where does the user experience factor into that because, you know, it's interesting. I, unfortunately, had an insurance claim recently with my car a deer hit me and...
KRISTINA: ..usually it's like I hit the deer but no, in this instance the deer hit me it actually ran out from the side of the road, and it hit the side of my car in the passenger side door and took off the mirror, but it's all on the passenger side so you can see it came, right? The site of the car, not the front in any way, was an interesting experience because what I could do is I could use my mobile device. I could upload photos of the damage and within, you know, 20 seconds. I had the approval of the claim. They said yes, you know $2000 - deductible, and you're on your way to find a reputable shop. And so I took it to a reputable shop, and they said, too much. It's more like 14 and a half thousand dollars. And so that's hugely different, and I'm wondering, you know, they had to have some kind of predictive analysis running. They had to be taking some kind of data points off of these photos that I uploaded for them, and they got it completely wrong, right, and so I went from being really excited that it was only 2,000 to oh gosh. It's more than 14,000, and they were great they covered it, but I didn't really think it was a great user experience. And so I'm wondering where is the user in the user experience and expectations factor into all of this in art, you know, am IA the anomaly is what I wanted to ask you or are we just in the early days of predictive analysis in the insurance business and it's going to take a few rungs to get it right?
AMIR: I think that mainly machine learning with predictive analytics, advanced analytics it's all synonyms for the same thing. It's very early in terms of usage, and I think people not exactly controlling the way it's being used correctly. And these explain most of you experience they are so eager to use it that they do not always remember how complicated is the mathematics behind it. And now how it is to be in the level of significance, which can move you into an autonomous process. So, if you - if we think about automatic, how we think like the car is automatic, but in fact when the car moves from manual to automatic because it isn't by default, I remember correctly. The driver was there it was just it was easy for him to do it. Moving to an autonomous car is a major change, but we all know what happened with the first test of autonomous cars, right? There was several incidents that won't support, and they're all based on aggravated that will not be tested to the ground. It's the same in insurance. So the moment you decide to move pauses into automatic or and then into autonomous, you need to be sure that you keep testing this algorithm; people usually with machine learning think that if we test it once like software, it will always work. In fact, with machine learning, that's not the case, machine learning performance and by performance, they may not automatically statistical performance the ability to predict usually deteriorate quickly like, nine months. It's unusable. And this can be another cause for that, so using this new tool requires a lot of understanding of the mathematics behind and also of the disadvantages and the need to maintain it all the time statistically. This explains what happen to you. The second issue is some of my lectures and talking about that. We used to believe in our data. I think that if we go back, we can hear, you know, really know why. And managers saying we have the greatest data. We only trust the data I have in my company, right? That's like a statement people might say, right you can hear someone saying that you've had it exactly. This is totally not the case because, for example, I would expect this insurance company to also collect a lot of data coming from the outside. For example, I will make a deal with some kind of a loud. I don't know. They count the fixing agencies and collect a lot of the transactions with pay the money for that. So they come on monetization buying data and they in data sharing is part of this business and having this is part of my model to feed my model to really understand what's going on. Not only my claims. So maybe if you are the largest insurance company in the world, maybe you can only feed you all waiting with your own data, but In most cases, it's much wiser to go share data and by data, and that's a typical example. I would buy all the invoices 10 years back and feed this they're celebrating. I am sure you will have a better user experience in this case.
KRISTINA: All right, so that that I should actually approach my insurance company and mention that to them, right? That's that's the next thing they should do it. So that brings me really to another question because I recently read that predicting the future is fun, like magic. And it's also essential to in the new economy, you know, you're heavily involved in the predictive analytics space. So I'm wondering if you can maybe speak a little bit about how come you know, if we look at some of the current surveys highlighting the power of predictive analytics, only 86 percent of organizations that use Predictive Analytics outperform their competition, right? It's not as high as I would think, and also, like 95 percent of managers say that they don't use predictive analytics, and they continue to make better decisions that aren't based on data. So how do we start to reconcile that instead of thinking about you know, you mentioned we're at the very early stages of ML, you know, need to do a lot of testing need to do a lot of data feeding and retesting if you will it what is sort of the perfect recipe or is there even a perfect recipe for a lot of these managers who have been resistant to do predictive analytics to start trusting and what are the guardrails we should be putting in place so that we can get it right? AMIR: Yes, it's a major issue, and I totally understand the statistics because, in a way, it's self-explanatory because people who are a little bit do not feel comfortable with, I would say, data from the future. That's what it means. In fact running machine learning. It's frightening. Rightmost of these people, the way to approach them is to run it in parallel, and what we usually do is tell them okay, we're not going now to replace anyone. That's not going to happen. Okay, no one is going to be replaced by robots. These are not robots with, you know, metalheads, etc. These are algorithms. Okay, but no one is going to be replaced by that but let's give you superpowers. There is a book called Human + Machine: Reimagining Work in the Age of AI and is using the term super-powers. I don't want to steal their amazing information. It's an amazing idea to put it that way. So when we superpower someone, we can show him. His own performance with and without it, and that's the best way to do it. And when you do it that way and you don't, you don't just assume that if some smart guy in the data division and the IT or even in the business, but more from the strategy. If this guy will come in and force the change on top of a business client, a business process, a major one that feeds the company, right? That's where the money comes from; there is a good chance that from the very beginning, it starts with him. I would say it should be treated as is a major company change to the company culture. The people are important in understanding, and you should do it in parallel. You should do it and always test yourself, and when you do it, you test the person who is going to make the decisions. So, you allow him to test himself to provide you say to test drive. And this new change and it's quite simple because it for example, if you think about time series put like forecasting of figures based on time to the future, and you want to plan your shifts. I don't know; you have between 50 to 500 people in a call center shift in some of the companies, right? These are massive numbers, but you need to decide how many people you need for the morning shift, the afternoon shift, in any of the weekdays, and we have holidays, and we have COVID which made everything very complicated, of course. So, when you do it, you take his own plan, and you show him: you go back you can do it very quickly because you can take his decisions from the past. Let's say from the last quarter every day that he did shift, today and you take the model prediction, and you show him that, if you accumulate everything and I'm talking about this case maybe they will like an I don't know, 1,200 men shifts that could be saved that were either or estimated by him for underestimated and when you look at the ways working today, you show him that there is no way that with just linear regression because this is not standalone. Okay, you could reach that. Because if you go into details and that's one of the things from the very beginning, another service is in business for it for individuals with for certified vehicles and data analytics services for 20 years. So, we always have been in a very detailed of the data the insurance data and one of the keys to machine learning are working with details a person a human cannot walk with the level of details that are not waiting can it's not only the mathematics which is inhuman. It's the level of data that you come from. So, when you take every piece of data, and you give the same attention to every piece of the transaction of traction in the call center on the shop floor immediately, you can see the fluctuations at the level of the afternoon on the last Wednesday, after the holidays and stuff like that. They can never do that. And when you collect these pieces, and you show him that's money and you become addicted and then what we spoke before if needs to be maintained because it can go very wrong within three months. If you don't really put attention constant attention to it.
KRISTINA: And it's actually a really important point. I think especially when we start looking at digital marketing and digital operations because there's no such thing. I mean if there is a snapshot in time, of course, but you know preferences are changing. We're in a situation where consumers are affected every day by little things. Right? Like I see a new ad or you know, my friend tells me about a new product all the way to wow. It's COVID my preferences have completely changed. So we really do have to continue to look at the new data and acquire data. I'm thinking about sort of this crazy universe we live in, and we're dealing with data and the need for data. Or to drive the algorithms to perfection to really help us out in business, but we're also at a point in time where data privacy matters more than ever. We've seen obviously impacts from GDPR, we've seen the CCPA, China is introducing its own data privacy regulation. Brazil's coming into effect. Next year South Africa is kind of doing it and has been doing it for a while. So, you know and there's 900 plus data privacy laws around the world. How do you balance that out, or how should we without the need for data privacy and yet the need for data not just to do a better job for the business but really to do a better job for consumers?
AMIR: It is a general rule: there is no issue of data privacy, which is connected to the issue of advanced analytics. Why, because any machine learning cannot learn only you solely; it needs to learn on a lot of people over a lot of time with many different interactions, with, for example, digital if that's the digital signatures that we are collecting. Okay, because we don't have a lot of digital signatures, the digital interface engagements provide us with many more digital signatures. When we look into that, a single person will never allow us to predict himself, and when you understand it, and I've been looking and working with GDPR-related issues for a long time, anything most of the regulations the arrive from each other in this way or another, the issue is not learning about a group of people. The issue is don't come back to me specifically. So, if you work with machine learning in a way where you eliminate the personality of the specific ID, you cannot go back to the same person. If you really look at the regulation and the idea behind the regulation. There is no issue with it. There is no issue in privacy in learning about people and implemented in a backup on paper. That's okay. It's good for both sides. The only thing is don't lean on me, and go back and use it against me specifically. And this is a good enough gap to have a lot of value. And the design and use the concept of obfuscation. I'm sure is it familiar. Okay. So, so, really the use of obfuscation to the depth allows you to ensure the regulator and yourself that you're not going to be doing any harm. If by accident, you will go back to similar people. That's okay. I didn't use the IDs. I didn't actually learn them. I've learned something about the community to the finest level, and this is fair enough. And I think this is what the European Regulators, at least which is the leader I think is looking for. So it's not it's not a barrier. I don't think people can say that because of privacy, machine learning cannot be pushed down the organization.
KRISTINA: Do you see there is a risk though there as well? So, for example, if a feeding data about individuals and removing those identifiers from the data, right, so that we're not talking or I'm not coming back to you consistently as an individual and doing repetitious data-gathering, what happens if I need to back out a piece of data because I find out that you know, it's either providing the wrong or skewed results or you know, for whatever reason I need to be able to go back and remove that data, you know. How do we deal with that then because now we've, you know pseudo-anonymized or even anonymize that data to a certain degree?
AMIR: Most of the anonymization obfuscation are reversible with the key. Now, this key should be kept very carefully, of course, but it can allow into the law because there's also one of the regulation requirements, so someone will be deselected, will be deleted, will disappear from the data set. And this can be achieved as they'll assuming this key is a one-way key that is kept in a secure place and never be used for long. I can tell you it's really a personal issue. I'm obsessed with data. I think I wrote on my LinkedIn. I really I'm like understand the benefits. I have an Android and iPhone, and I have like, I think I'm not exaggerating now. I think I have like 250 applications on my Android, and I'm connected to 12 social networks and many others. So, I feel okay with that with major suppliers. And I know that what is being done will not harm me if because if it's like it's a top-notch supplier, and I also know that you need to give something to get something, and I dramatically enjoy the benefit of that and not saying it should be open and just saying we should not be scared of it and we should make sure that people that are scared of it can be opt-out from that and also if they can decide if they want to decide to do it afterward. I can tell you, for example, with one of the insurance companies that I work with, they said okay, we want to ask the customer to share the data between companies in the group because legally, it was if you have consent with one of the companies that do a lot M&As in this, you know that so, so, this cover should be done in the group. Okay. So the estimation was that five percent of the people would agree. I think like seventy percent agreed. Why? Because it was presented in the customer experience way of what will you get next because you can do cross-product between life and P&C and this is good for you. Maybe we can offer you something better and it was explained. So more than 50 percent agreed to do that out of the blue without getting money or because they had to do it in some kind of a conflict when they signed in it was the post-mortem it was done from the marketing side, but explained in this specific concept were explaining. What are we doing with the data? And this was like two years ago. So I think now the situation is much better and COVID I think, opened everything, right?
KRISTINA: COVID changed everything for everyone, and it hopefully in a good way, right?
AMIR: I honestly believe so. Of course, people out there then and sick then and not didn't know this part but generally speaking, the changes we were waiting for and people were pitching and then were like, "oh these are amazing ideas", but "not here, not now, maybe later". The customer will give us his email it's private stuff like you don't hear it anymore. It's finished immediately, even with double opt-in. I don't know exactly how is it think they say that the state's level in the States. It's, but I can tell you that in Israel. There was no it was a process of making people using the government digital, etc. COVID started how it started. This is ended immediately. Like I think 85% of the people have a digital account and they get services that they're being paid. And it's helping. This happened in two months. Let's to see it. It was amazing.
KRISTINA: I don't mind having my data collected and used, but I really want something in return, and it better be valuable. It doesn't have to be money. In fact, usually, it's not money. It's usually convenience. It's my time. It's something else, and so that is something that I value, and I put a premium on, and so, you know if we're going down this path, it just really seems like predictive analytics data is really where it's at, not just like for this year, but for this decade as I started saying, right and for decades to come probably?
AMIR: And I think we can see it in some of the other companies and in fact what you should we always limit should remember is that customer experience is about having the engagement when you need it and not too much. In fact, we don't really want to talk to our insurance company, right? We wanted it to be ahead of us. And when it's something which needs to be done now, we need them to know everything on the spot and we need we want them to predict something that can be done in advance. So, if you combine all of these, you can always technologies that are available, data that is available maybe not in the company, but it's definitely available to someone in a digital manner. And if you are an insurance company with ensuring farmers somewhere in the deep South of the United States, they can even warn the farmers ahead of time because there is some kind of a catastrophe that is coming in. Right, it's a catastrophe. It comes from French because I speak French, so sorry. So, just imagine if you are a farmer. You don't have the best weather forecasting that is also looking at the type of crop you have, etc. And you get a warning that your crop might be in danger, do something today, the insurance company has the means to give you that we show this capability to insurance companies and everything is available. The only thing you need to do is to think out of the box, as I'm the insurance company and I am waiting for your claim. No! Why not warn me in advance? And by the way, you mentioned the deer hitting your car, which is a bit like the man that was biting the dog. Yeah. The same the same joke. So is this old known for being something like that that the animals of the road?
KRISTINA: Absolutely. I mean, you know, it.
AMIR: Exactly. I was expecting that!
KRISTINA: Absolutely. In fact, I was just talking to a friend of mine. Who said she moved closer to where I live and because she's moved here. Yeah, I think it's like I don't know how many miles, but it's something like 20 kilometers closer to me. This area actually has more deer because it's Development Area used to be a forest. There's lots of deer, and they are charging a premium, right, or a higher premium because they know there's more deer on the road. So this isn't anything that's a news flash to anybody. In fact, they expect it to happen. So it was predictable to some extent. Yeah. It was just a matter of how many people Would be hit by a deer.
AMIR: and most of the most what time of the day was it is it time of the day like the these are expected to be out maybe the reason I don't know. I don't know the details.
AMIR: So why not get the warning, like Waze? They know you. You have their app on your phone. Why did you not get a warning? And I'm sure they have a claim from this specific piece of the road, and I'm sure they could have gotten accident reports or claims reports from other sources, which should help give you a useful service. This is a complete change of the game because the moment insurance companies will have more data, and they can, I would say in use it themselves. They even understand they can do much more. It is called data monetization, which is a bit of far-fetched for most of the insurance industry, but it's starting.
KRISTINA: So that's fascinating. Okay. This is just an aside, but I'm going to say it to you because I'm actually going to go take a photo of this for you. There is a very interesting road sign that a neighbor of mine put up on his fence and it says, warning deer crossing. Right. And that sign has been there for at least 20 years if not more, so my neighbor knows that dears crossed there.
AMIR: But a multi-billion insurance company doesn't, right? Exactly.
KRISTINA: Okay. Well, you know what? You've made me excited because of the possibilities, or you made me more excited about the possibilities, but you've also made me realize just how much of a gap there is, right? And so, I think the challenge is for all of us to step up to the plate and to get more energized and excited about what we can do if we think outside of the box because, as you said, the data is available. We're starting to have the mechanisms, and yes, still early on in machine learning, but you know what, no time like the present.
AMIR: And technology is real. All of these are now the issue is machine learning and databases and user interface. No technology issued, or it's finished. It's there it's available. But I was saying about machine learning is how to use it. It's not the technology; the technology is available. So it is there is some kind of a barrier that is not really it's more about I think who's that will be the first to move in your competitive environment if your insurance company and X or Y.
KRISTINA: I think that you know for me, it's always hard to say goodbye, especially when we're talking about such an interesting topic, but I think that's a really good challenge to leave with our listeners, which is who's going to be the first to move and it could be you, and it could be a competitive advantage right? And so you could realize a lot of opportunities, but you can't stand by and just look at the risk and do nothing. It's really a time to become active, and you can jump in with two feet, and you're still not going to be jumping too far ahead. Are you getting ahead of yourself? So Amir, thank you for taking the time to join me today and to everyone listening. Thank you for hanging out with us. I wish we had another half an hour at least to talk about all of these topics. But until next time, be well do powerful digital policy work and Amir. Thank you again for taking the time to be with us today.
AMIR: Thank you very much exciting conversation.