AI Transformation: Unveiling its Generative Power with Microsoft and OpenAI
Insurtech has been revolutionized by advancements in AI technology, and in particular, Gen AI, which continues to transform underwriting, customer service, fraud detection, and more. The partnership between market leader Microsoft and groundbreaking AI R&D company, OpenAI, has helped bring innovative Generative AI models to market and supports greater research into how these technologies can benefit business and society. Special guest Ben Haklai, National Technology Officer for Microsoft, joins Sapiens Chief Strategy Officer, Alex Zukerman, to examine this important partnership and how it continues to move Gen AI and its applications forward in our latest podcast.
Alex Zukerman|Ben Haklai
Alex Zukerman: Hello! Welcome the Sapiens Insurance 360 podcast. I’m your host, Alex Zuckerman, Chief Strategy Officer at Sapiens, and I’m so glad that you’re out there listening. This is where we discuss the latest news, trends, and issues from across the insurance solutions and technology spectrum.
One of the reasons that I love working in insurtech is that its landscape is constantly evolving with new and innovative technologies, which constantly raise the bar in terms of next-generation functionality. Today, generative AI is the technology that is driving the insurtech evolution, and it transforms such areas as underwriting the claims cycle, customer service, marketing fraud detection, and much more. And the key partnership that is shaping the future of generative AI solutions is the one between Microsoft and OpenAI. In today’s podcast, “AI Transformation: Unveiling its Generative Power with Microsoft and OpenAI,” we’ll discuss this all-important partnership and how it continues to move Generative AI and its applications forward. Our special guest today is Ben Haklai, National Technology Officer for Microsoft, who sheds some light on this exciting partnership and how it continues to shape the future of generative AI solutions. In his role, Ben partners with organizations, governments, and regulators to help them realize the promise of new technologies while mitigating regulatory and legal risks. His unique skill set includes a deep understanding of AI, AI ethics, privacy, cybersecurity, and societal change, and how these areas drive and impact us all. Ben, welcome to the program!
Ben Haklai: Thanks, Alex. That’s quite the introduction and I’m happy to be here. Thank you for having me!
Alex Zukerman: Absolutely, Ben. So to start with, at the very beginning, could you please elaborate on [the] Microsoft partnership with OpenAI and how it contributes to the advancing of AI technology and specifically, Gen AI technology?
Ben Haklai: Absolutely. So, Microsoft or AI is not new to Microsoft. We’ve been a market leader for the last 15 and even more years, in various areas. But obviously this unique partnership with OpenAI has been a huge one for us and brought us to the forefront of the Gen AI race. So I won’t go into the details of everything that you [already] know, but it’s a partnership that goes back to 2017, and even before that, the first big investment, 2019, was when this partnership actually consolidated into something very tangible, and allowed OpenAI to practically use the compute resources that Microsoft has in Azure and allowed us to partner in bringing the most innovative models to the market. So that’s the GPT family of models, and obviously, the big boom, 2021 ChatGPT, which was originally based on GPT 3.5 came to the market, became the most, or the fastest adopted, adopted technology in history, in humankind history, and, you know, brought this technology to the fingertips of us all. So that’s quite unique.
Alex Zukerman: Great, great, great intro, Ben, thank you. To a big question that is in the headlines everywhere and definitely in the insurance industry as a, regulated industry with a unique responsibility toward its customers, and the question of security, data privacy, bias and all those aspects to our becoming or, consider it as a concern global concern on the generative AI. So can you give us a bit details on, your view on that and what Microsoft [is] doing to solve this challenge?
Ben Haklai: Absolutely. So as we partnered with OpenAI, that partnership became closer and closer, but still, it’s two different companies, right? So we have Microsoft, we have OpenAI. And for us, it was very clear that for us to be able to succeed in this market, we have to work with our closest customers, listen to their concerns, and try to cater for those concerns as we build those solutions. So it took us a while longer to get to the market with some of the solutions, although we partner in building some of those GPT models, mainly due to the reason that we had to bake in, a lot of, tools and capabilities in order to mitigate some of the risks you mentioned. So when we talk to our customers, so they raised data and security, right? Responsible AI compliance, and intellectual property as top concerns that they have. And we built our solutions with that in mind, in order to try and mitigate many of the risks that our customers might face as they start this journey. And again, obviously you talked about insurance. Insurance is obviously one of those industries where regulation plays a big part, where risk plays a big part. And again, we wanted to build the models and the services in a way that the data remains our customers’ data. It does not train the model. We keep the competitive edge of our partners and customers in their, you know, goal to improve their businesses. We want to make sure we reduce bias. We reduce hallucinations in the models. So we have a lot of filtering layers. We build the prompt in a different way. We have meta prompts in the background. Our solutions meet the most rigid compliance criteria in the market. So when you use Azure, OpenAI, or [a] Copilot stack, you actually enjoy all of the certification that Azure meets out of the box. And then last on the intellectual property piece, we actually, will protect our customers, in case that they’re sued out of, response that they get from the model. So if I have a customer who uses our services and gets sued in court by someone else, we, Microsoft, will manage that lawsuit, and will pay any adverse outcome. So we try again to bake in and build, you know, a sort of a comprehensive solution to meet our customers’ needs in specific industries. Insurance is obviously one of those.
Alex Zukerman: This is great, I’m sure that our listeners will be super excited about that because again, as we said, this is definitely a point that the industry is looking how, how companies can solve it, so thank you for that.
Ben Haklai: Absolutely, absolutely. And again, that’s a huge concern. One of many, right? We also need to make sure we stay at the forefront of the technology development. Our customers are very demanding. They want to use the best in class. So we are trying to offer a platform that goes end to end, offering the biggest, most advanced models, GPTs, for example, on the one side, and small language models, SLM on the others.
So on the other side, so our customers will be able to cater for each one of their needs. Sometimes you want to use an SLM and not that large language model for a specific use case, right? You want to drive the cost to be lower, etc., etc. So we want to offer that richness in the platform, so our customers can do whatever they want and desire to do and drive productivity.
Alex Zukerman: And this is a great approach. And I think in general, too, when it’s something that we also talked about previously offline is together with all the risk. We need not to forget the huge, huge benefits, advantages and innovation that those capabilities bring. And if we look at the insurance industry, which is, very, very heavy on documentation, on text, on correspondence, on documents, and spends, and even though that all of them are digitalized today and are running through digital channels, they’ll still, involving a lot of a human, analysis, a lot of a human labor in order to analyze, review, and make effect of all those documents. So capabilities such as NLM that really is a copilot on steroids, not replacing the users, not changing the role, but actually providing a huge support, to those users who work across the insurance value chain and provides them a tool on steroids to perform the work in a much higher efficiency, in a much faster rate, and to come to a better result. If we look at the most important things on the agenda of CIOs today, is how do we protect ourselves from cybersecurity and how we achieve better operational efficiency. Operational efficiency in this volatile, ever-changing, financially unstable market is a critical factor today to insurance companies, and what you describe and the use cases that we are starting to see and to build into the insurance industry, and across the value chain brings a huge amount of operational efficiency, of speed, of reduction of effort, and better and more accurate results across the value chain.
And when we talk about those use cases, what we need to remember, it’s not only for the users of the insurance across all claim handlers, underwriters, agents, customer service, etc. This is only one aspect of the of the triangle. The two other parts of of this equation are the back-office users. So how do we configure a system so the ability to automate configuration, to automate change, and to drive change in the business quickly using NLM is super powerful. And then expanding it yet to another area of training, of knowledge sharing, and of ability to automate processes like document creation, like data migration. So implement faster, run your business faster, and give a better service. So really across the board, those capabilities can bring huge efficiency. And again, I really like the way that Satya Nadella, Microsoft CEO, when he describes those capabilities, they really focus on the copilot aspect of that. And again, it’s not, removing the way insurance works, but just improving it by, by a huge factor.
Ben Haklai: Absolutely. And you talked about the labor-intensive work and the ability to digest huge amounts of documents and reason over them. I think that’s one of the areas these models really shine in, right? So they’re able to digest huge amounts of unstructured data, reason over it, and generate a lot of value for you as the end customer or the data owner or your end customers, in order to drive efficiencies, in order to drive better results, in order to drive frictionless engagement with an end customer. All of those are baked into those services and allow you to kind of, look to the future in a very optimistic way.
Alex Zukerman: And an optimistic look on the future is really bright. And this is what we want to have here. So Ben, as a last question, and I think it’s a big one. So, if you want to take a big leap forward and think how Gen AI specifically will influence not only the insurance industry, but actually the society or enterprise, behavior and commercial behavior, can you please share your thoughts on that?
Ben Haklai: So the first thing I’ll say is that we actually believe that Gen AI will influence every industry and at the end of the day, each one of us, in a very productive way, it will give us, a new assistant or, you know, an ability to do more than we are doing today. And when we look to the future and, again, I’ll take something that Jensen Huang, the CEO of Nvidia, said recently, we might get to a place where compute costs are marginal, right? We see an increase in compute capacity every year. Microsoft invests north of $30 billion a year into [building] that infrastructure. And we see an increase of 30x in compute raw power. So practically, at the end of the day, we’ll get to a point where the compute cost is driven so low that it becomes marginal. When we get to that point, we’ll see models that keep training on their own, inferencing all at the same time. They will hallucinate or come up with ideas, and then we’ll be able to ground them in reality as they actually experience everything around us. So we’ll get to a point where, you know, we are those models will become much more able than they are today. They’ll be able to reason in a better capacity and contribute more to society as a whole. We’ll see a lot of, you know, very, very positive influences in across industries, in healthcare, in insurance, in banking, in governments, all over the place. So that’s what we hope to see in the very near future.
Alex Zukerman: Thanks a lot for the perspective. And then I think the emergence of the generative AI technology and the impact of [it] on the insurance marketplace and the insurance industry is yet to be understood, but it’s surely going to be one of the major, major, influencers on what we do in this industry, even thinking on solution providers, what we see already in the short term, is a lot of capabilities that were, there are a lot of startups that were created to deal with a specific need, whether it’s translate PDF to word, whether it’s transform [one] format to another, analyze, fix, etc. Suddenly a huge part of this world becomes redundant because the technology is base and is ready to be used by everybody. So this is, we start to see it now in the short looking on the longer term. This view of a personal assistant that is trained for you, that knows your way, a small database that runs with you and can enable to help you in all your activities, including your work activities on the insurance. I think this is an amazing vision, that will really take humanity even to a different place, but even if you look on the technology side in the midterm, how much do we need to create dashboards upfront? How [many] templates do we need when we have the ability to analyze the ad doc text data and get the specific results? For us, tailored for a specific case and not in the general level? I think this is also already quite a revolution, and it’s going to be super, super exciting to see what’s coming next.
Ben Haklai: Absolutely.
Alex Zukerman: So to our listeners, as always, thank you very much for spending your time with us here today. We love hearing from you, so if you have comments, suggestions or would like to follow us on social media, please reach out to us on our channels and don’t forget to subscribe to the podcast and thank you once again for listening. We’ve got more coming, so be sure to tune in the next time to Sapiens Insurance 360. Thank you very much.
